Enzymatic and structural characterization of hydrolysis of gibberellin A4 glucosyl ester by a rice β-D-glucosidase

Article abstract of DOI:10.1016/j.abb.2013.06.005

In order to identify a rice gibberellin ester b-D-glucosidase, gibberellin A4 β-D-glucosyl ester (GA₄-GE) was synthesized and used to screen rice β-Glucosidases. Os3BGlu6 was found to have the highest hydrolysis activity to GA₄-GE among five recombinantly expressed rice glycoside hydrolase family GH1 enzymes from different phylogenic clusters. The kinetic parameters of Os3BGlu6 and its mutants E178Q, E178A, E394D, E394Q and M251N for hydrolysis of p-nitrophenyl b-D-glucopyranoside (pNPGlc) and GA ₄-GE confirmed the roles of the catalytic acid/base and nucleophile for hydrolysis of both substrates and suggested M251 contributes to binding hydrophobic aglycones. The activities of the Os3BGlu6 E178Q and E178A acid/base mutants were rescued by azide, which they transglucosylate to produce b-D-glucopyranosyl azide, in a pH-dependent manner, while acetate also rescued Os3BGlu6 E178A at low pH. High concentrations of sodium azide (200-400 mM) inhibited Os3BGlu6 E178Q but not Os3BGlu6 E178A. The structures of Os3BGlu6 E178Q crystallized with either GA₄-GE or pNPGlc had a native a-D-glucosyl moiety covalently linked to the catalytic nucleophile, E394, which showed the hydrogen bonding to the 2-hydroxyl in the covalent intermediate. These data suggest that a GH1 β-Glucosidase uses the same retaining catalytic mechanism to hydrolyze 1-O-acyl glucose ester and glucoside.

Full text of DOI:10.1016/j.abb.2013.06.005

Archives of Biochemistry and Biophysics 537 (2013) 39–48
Contents lists available at SciVerse ScienceDirect
Archives of Biochemistry and Biophysics
journal homepage: www.elsevier.com/locate/yabbi
Enzymatic and structural characterization of hydrolysis of gibberellin A4
glucosyl ester by a rice b-D-glucosidase
Yanling Hua a,b, Sompong Sansenya a, Chiraporn Saetang a, Shinji Wakuta a,1, James R. Ketudat Cairns a,
⇑
a School of Biochemistry and Chemistry, Institute of Science, Suranaree University of Technology, Nakhon Ratchasima 30000, Thailand
b The Center for Scientific and Technological Equipment, Suranaree University of Technology, Nakhon Ratchasima 30000, Thailand
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 15 May 2013
and in revised form 11 June 2013
Available online 26 June 2013
In order to identify a rice gibberellin ester b-
D
-glucosidase, gibberellin A4 b-
D
-glucosyl ester (GA4-GE) was
synthesized and used to screen rice b-glucosidases. Os3BGlu6 was found to have the highest hydrolysis
activity to GA4-GE among five recombinantly expressed rice glycoside hydrolase family GH1 enzymes
from different phylogenic clusters. The kinetic parameters of Os3BGlu6 and its mutants E178Q, E178A,
E394D, E394Q and M251N for hydrolysis of p-nitrophenyl b-D-glucopyranoside (pNPGlc) and GA4-GE
confirmed the roles of the catalytic acid/base and nucleophile for hydrolysis of both substrates and sug-
gested M251 contributes to binding hydrophobic aglycones. The activities of the Os3BGlu6 E178Q and
E178A acid/base mutants were rescued by azide, which they transglucosylate to produce b-D-glucopyr-
anosyl azide, in a pH-dependent manner, while acetate also rescued Os3BGlu6 E178A at low pH. High
concentrations of sodium azide (200–400 mM) inhibited Os3BGlu6 E178Q but not Os3BGlu6 E178A.
Keywords:
b-Glucosidase
Gibberellin glucosyl ester
Covalent intermediate
X-ray crystal structure
Transglycosylation
The structures of Os3BGlu6 E178Q crystallized with either GA4-GE or pNPGlc had a native a-D-glucosyl
moiety covalently linked to the catalytic nucleophile, E394, which showed the hydrogen bonding to
the 2-hydroxyl in the covalent intermediate. These data suggest that a GH1 b-glucosidase uses the same
retaining catalytic mechanism to hydrolyze 1-O-acyl glucose ester and glucoside.
Ó 2013 Elsevier Inc. All rights reserved.
Introduction
b-Glucosidases have been classified into glycoside hydrolase
(GH)2 families GH1, GH3, GH5, GH9, GH30, and GH116, based on
Beta-glucosidases (b-
3.2.1.21) are enzymes that hydrolyze the b-O-glycosidic bond at
the anomeric carbon of a nonreducing terminal -glucosyl moiety
to release -glucose from glycosides and oligosaccharides. These
D
-glucopyranoside glucohydrolases, E.C.
their amino acid sequences and structural similarity, while other
b-glucosidases remain to be classified [9–12]. Most characterized
plant b-glucosidases belong to GH1, which falls in GH Clan A, as
D
D
do GH5 and GH30. GH Clan A enzymes have (b/a)8-barrel structures
enzymes are found in essentially all living organisms and have
been implicated in a diversity of roles, such as biomass conversion
in microorganisms [1] and activation of defense compounds [2,3],
phytohormones [4,5], lignin precursors [6], aromatic volatiles [7],
and metabolic intermediates [8] by releasing glucose blocking
groups from their inactive glucosides in plants. To achieve specific-
ity for these various functions, different b-glucosidases must bind
to distinct aglycones with a wide variety of structures, in addition
to the glucose of the substrate.
with catalytic acid/base and nucleophile on the ends of strands 4 and
7 of the barrel, respectively. The mechanism by which GH1 enzymes
recognize and hydrolyze substrates with different specificities re-
mains an area of intense study.
Among b-glucosidase functions in plants, the function of phyto-
hormone activation has attracted many peoples’ interest. Phyto-
hormone glucosyl conjugates are ubiquitous in plants, including
those of gibberellic acids (GAs), abscisic acid (ABA), cytokinins,
2
Abbreviations used: A405, absorbance at 405 nm; ABA, abscisic acid; ABA-GE, ABA
b-D-glucosyl ester; API-ES, atmospheric pressure ionization-electrospray; DAD, diode
array detector; DMSO, dimethyl sulfoxide; DNP2FG, 2,4-dinitrophenyl-2-deoxy-2-
fluoro-b-D-glucopyranoside; EtOAc, ethyl acetate; GA, gibberellic acid(s); GA4-GE,
gibberellin A4 b-D-glucosyl ester; G2F, 2-deoxy-2-fluoroglucoside; GH, glycoside
hydrolase; IAA, indole acetic acid; IAA-GE, IAA b-D-glucosyl ester; IMAC, immobilized
metal affinity chromatography; MeOH, Methanol; MES, 2-(N-morpholino)ethanesul-
fonic acid; NaOAc, sodium acetate; PEG5000MME, polyethylene glycol 5000 mono-
methyl ester; pNPG, p-nitrophenol; pNPGlc, p-nitrophenyl b-D-glucopyranoside; TEV,
tobacco etch virus (protease).
⇑
Corresponding author. Address: Institute of Science, Suranaree University of
Technology, 111 University Avenue, Nakhon Ratchasima 30000, Thailand. Fax: +66
44 224185.
E-mail address: cairns@sut.ac.th (J.R. Ketudat Cairns).
Present address: Research Faculty of Agriculture, Hokkaido University, Kita-ku,
1
Sapporo 062-8555, Japan.
0003-9861/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.abb.2013.06.005

40
Y. Hua et al. / Archives of Biochemistry and Biophysics 537 (2013) 39–48
auxins, and jasmonic acid derivatives, and b-glucosidases that
hydrolyze these conjugates have been found in plants [4,13–20].
ABA and the auxin indole-3-acetic acid (IAA) have both been
found as inactive 1-O-acyl glucose esters in plants. ABA is critical
for plant growth, development, and adaptation to various stress
conditions. Plants have to adjust ABA levels constantly to respond
to changing physiological and environmental conditions [14]. Ab-
scisic acid glucose ester (ABA-GE) is a biologically inactive form
that constitutes a reserve form of ABA. Two Arabidopsis b-glucosi-
dase (AtBG1 and AtBG2) were found to hydrolyze ABA-GE to form
free ABA and to be essential to proper response to drought stress
and delay in seed germination, despite the fact that other ABA-
GE hydrolyzing enzymes were detected [14,17]. IAA glucosyl ester
(IAA-GE) is also an inactive, stored form that can also be hydro-
lyzed by IAA-glucose hydrolase to release active IAA [15].
we detected [M+Na]+ at m/z 517.1, [M+H–H2O]+ at m/z 477.2,
and [M+H–Glc]+ at m/z 315.5.
Expression of rice GH1 enzymes to screen for GA4-GE hydrolysis
Five GH1 enzymes that have been expressed in our laboratory,
Os3BGlu6 [26], Os3BGlu7 [27], Os4BGlu12 [28], Os4BGlu18 (S. Bai-
ya et al., unpublished) and Os9BGlu31 [29] were expressed as N-
terminal thioredoxin fusion proteins in Escherichia coli and purified
by immobilized metal affinity chromatography, as previously de-
scribed. The purified fusion proteins were tested for the hydrolysis
activity to pNPGlc and GA4-GE according to the method described
below.
Site-directed mutagenesis of Os3BGlu6
Both 1-O-acyl glucosyl esters and glucosides of GAs are found in
plants. GAs promote germination, shoot elongation, and flower
development, among their many functions [21]. It has been sug-
gested that GA glucosyl esters are deactivated GAs that can be enzy-
matically reconverted to active GAs, thus serving as a reserve form of
Mutagenesis of the pET32/Os3BGlu6 expression vector [26] to
create the Os3BGlu6E178A, Os3BGlu6E178Q, Os3BGlu6E394D
and Os3BGlu6E394Q mutations was performed with the Quik-
ChangeÒ Site-Directed Mutagenesis Kit (Stratagene, Agilent Corp.,
La Jolla, CA, USA) according to the supplier’s instructions. The fol-
lowing oligonucleotides were used for mutagenesis: for E178A,
50GATCACGCTCAACGCGCCGCACACGGTG30 and its reverse comple-
ment; for E178Q, 50GGATCACGCTCAACCAACCGCACACGGTGGC30
and its reverse complement; for E394D, 50CCAGTGTACATCACTG
ATAACGGGATGGATGACAGC30 and its reverse complement; and
for E394Q, 50CCACCAGTGTACATCACTCAGAACGGGATGGATGA-
CAGC30 and its reverse complement. The cDNA were confirmed
to include the desired mutations and be free of additional muta-
tions by automated DNA sequencing (Macrogen Corp., Rep. of
Korea).
biologically active GAs [22]. After [13C]GA20-b-
D
-glucosyl ester was
injected into light-grown maize seedlings, the metabolites,
13C]GA20, [13C]GA29, [13C]GA20-13-O-glucoside, [13C]GA29-2-O-glu-
[
coside, [13C]GA8 and [13C]GA8-2-O-glucoside were identified in the
extracts of the seedlings made 24 h after the injection [23]. This
showed that the endogenous hydrolysis of the introduced conjugate
and its reconjugation led to the three new glucosides. In rice,
[3H]GA1, [3H]GA2, [3H]GA34, the glucosides of [3H]GA2, [3H]GA4,
[3H]GA8 and [3H]GA34, and the glucosyl ester of [3H]GA4 (GA4-GE)
have been found after application of [3H]GA4 to cell suspension cul-
tures of Oryza sativa cv. nipponbare [24].
b-Glucosidases have been proposed to hydrolyze the GA conju-
gates to active GAs, but the molecular identification of these en-
zymes and investigation of their modes of action have yet to be
reported. Schliemann [13] reported that b-glucosidases extracted
from mature rice seeds and seedlings have different hydrolytic
activities toward GA8-2-O-glucoside, GA3-3-O-glucoside and 1-O-
GA3-glucosyl ester, but he did not purify and characterize the b-
glucosidases. Furthermore, the substrate specificity of b-glucosi-
dase to GA conjugates and how b-glucosidase binds to GA conju-
gates has yet to be reported. In this study, we identified a rice
Recombinant expression and purification of the mutants of Os3BGlu6
The wild type rice Os3BGlu6 and its mutants M251N, E178Q,
E178A, E394D and E394Q were expressed in E. coli strain Origami
(DE3) as fusion proteins with N-terminal thioredoxin and His6 tags
as described previously for Os3BGlu6 [26]. The crude proteins were
first purified by immobilized metal (Co2+) affinity chromatography
(IMAC). The N-terminal thioredoxin, His6 and S tags were then ex-
cised with TEV protease, and removed with a second IMAC column
purification. Protein concentrations were estimated by the Brad-
ford protein assay (Bio-Rad) with bovine serum albumin as the
standard. The values from this assay were in the good agreement
with those obtained from the 280 nm absorbance with the calcu-
lated extinction coefficient.
GA4-GE b-D-glucosidase and characterized its mechanism of glu-
cosyl ester hydrolysis by comparing hydrolysis and transglycosy-
lation activities of wild type and acid/base mutant enzymes with
GA4-GE. To better understand the ester-enzyme interaction, struc-
tures of mutant enzymes soaked with GA4-GE and pNPGlc were
also determined.
Determination of pH optimum for Os3BGlu6 and its mutants
Materials and methods
The optimum pH of Os3BGlu6 and Os3BGlu6 M251N hydrolysis
of pNPGlc were determined by incubating 1
lg of enzyme with
Synthesis of GA4-GE
2 mM pNPGlc in 80 l of 100 mM universal buffer (citric acid-diso-
l
dium hydrogen phosphate), pH 2.0–11.0 in 0.5-pH-unit incre-
GA4-GE was synthesized from GA4 (Jiangsu Fengyuan Bioengi-
neering Co.Ltd, P.R. China) following the method of Hiraga, et al.
[25]. The acetylated and deacetylated GA4-GE were obtained with
43.7% and 40.5% yields, respectively. The synthesized acetylated
and deacetylated GA4-GE structures were confirmed by NMR spec-
tra on a 300 MHz NMR spectrometer with a Varian 300 ID/PFG
probe at a frequency of 299.986 MHz (Unity INOVA, Varian, USA).
Deuterated chloroform (CDCl3) and dimethyl sulfoxide-d6
(DMSO-d6) were used as solvents for acetylated and deacetylated
GA4-GE, respectively. The 1H NMR was consistent with the pub-
lished data for GA4-GE [25].
ments, at 30 °C for 10 min. The reactions were stopped by adding
100 ll of 2 M sodium carbonate. The released p-nitrophenol
(pNP) was quantified by measuring the absorbance at 405 nm
(A405) with a microplate reader (Thermo Labsystems, Finland)
and comparing it to that of a pNP standard curve.
The optimum pH of Os3BGlu6 and its mutants M251N, E178Q
and E178A for hydrolysis of GA4-GE were determined by incubat-
ing 1
lg of Os3BGlu6 or Os3BGlu6 M251N, or 5.0
lg of E178Q or
E178A with 0.86 mM GA4-GE in 80
ll of 100 mM universal buffer,
pH 2.0–11.0 in 0.5-pH-unit increments, at 30° for 20 min. The reac-
tions were stopped by boiling 1 min and cooled on ice immedi-
ately. The released glucose was quantified with a glucose oxidase
assay [27].
The identity of the deacetylated GA4-GE was also confirmed
from its mass spectrum (data not shown). In the positive mode,

Y. Hua et al. / Archives of Biochemistry and Biophysics 537 (2013) 39–48
41
Measurement of wild type and mutant Os3BGlu6 activities for
hydrolysis of GA4-GE and pNPGlc
formed in 50 mM MES, pH 5, at 30 °C for 20 min. The turnover rates
of GA4 release per minute per unit enzyme (V0/E0) were calculated
based on the GA4 peak area. For the reactions varying the concen-
trations of GA4-GE donor, sodium azide was fixed at 50, 100, 200
and 400 mM, and compared to reactions without sodium azide.
The activities of Os3BGlu6 wild type and mutants to hydrolyze
pNPGlc were measured as described for the pH optimum determi-
nation, but in 50 mM sodium acetate (NaOAc) buffer, pH 5.0. To
measure hydrolysis activities toward GA4-GE, Os3BGlu6 wild type
and M251N mutant enzymes were incubated with GA4-GE in
50 mM NaOAc, pH 5.0, but 50 mM 2-(N-morpholino)ethanesulfo-
nic acid (MES) buffer was selected for the E178Q and E178A mu-
tants to avoid transglucosylation. For determination of kinetic
parameters, variable reaction times, enzyme amounts and sub-
strate concentrations were tested to obtain the initial velocities.
The Km and kcat were calculated from nonlinear regression of
Michaelis–Menten plots with Grafit 5.0 software (Erithacus Soft-
ware, Horley, UK).
Crystallization of Os3BGlu6 E178Q
Purified Os3BGlu6 E178Q and Os3BGlu6 E394D were concen-
trated to 5 mg/ml and crystallized by hanging drop vapor diffusion
with a precipitant of 13% polyethylene glycol 5000 monomethyl
ester (PEG5000MME), 0.1 M Bis/Tris, pH 6.5, for Os3BGlu6 E178Q
with pNPGlc and Os3BGlu6 E178Q with GA4-GE, and 17%
PEG5000MME, 0.1 M Bis/Tris, pH 6.5 for Os3BGlu6 E394D with
GA4-GE. The crystals were approximately 0.5 ꢀ 0.1 ꢀ 0.1 mm in
size after growing 2 days. For cryo-protection, the crystals were
transferred to a solution consisting of 22% PEG5000MME, 0.2 M
Bis/Tris, pH 6.5, 18% (v/v) glycerol, which contained 10 mM GA4-
GE or pNPGlc.
Identification of transglucosylation product with TLC, LC-MS and NMR
Transglucosylation reactions were studied for Os3BGlu6 wild
type, M251N, E178Q and E178A mutants. 2 mM pNPGlc, 5 mM
X-ray diffraction data collection and processing
GA4-GE or 2,4-dinitrophenyl-2-deoxy-2-fluoro-b-
side (DNP2FG) donor was reacted with 1.0 g of enzymes in
50 mM MES buffer, pH 5; and 40 mM sodium azide or 0.96 mM
free GA4 was used as acceptor. At different times, 10 l aliquots
D-glucopyrano-
l
The data sets were collected at 100 K in a nitrogen stream
(CryoJet; Oxford Instrument) on the PX 13B beamline at the Na-
tional Synchrotron Radiation Research Center, Hsinchu, Taiwan,
with a 1.0 Å wavelength X-ray beam and an ADSC Quantum 315
CCD detector. All the diffraction images were indexed, integrated
and scaled with the HKL-2000 program [30]. The crystals of Os3B-
Glu6 E178Q soaked with GA4-GE and pNPGlc belonged to the
orthorhombic space group P212121 and diffracted to 1.97 and
1.90 Å resolutions, respectively. Unit cell parameters for Os3BGlu6
E178Q with GA4-GE were a = 57.1, b = 91.2, c = 111.4 Å,
l
of each reaction mix were removed, boiled 1 min, and kept on ice
for TLC analysis. Reactions without enzymes were used as controls.
Sample aliquots were spotted on TLC plates, which were developed
with EtOAc–MeOH–H2O (7.5:2.5:1.0) or CHCl3–MeOH (7:3). The
products were detected by staining with 10% sulfuric acid in etha-
nol followed by charring.
The new transglucosylation products of E178Q and E178A were
analyzed by LC-MS. Samples were separated over a ZORBAX Eclipse
a
= b =
c = 111.2 Å,
a = 56.5, b = 91.2, c = 111.2 Å,
c
= 90°, while those with pNPGlc were a = 57.0, b = 91.1,
= b = = 90° and for Os3BGlu6 E394D_GA4-GE were
= b = = 90°. Data-collection and
XDB-C18 column (4.6 ꢀ 150 mm, 5
lm, Agilent, USA) on an Agilent
a
c
1100 HPLC. A gradient of 0–80% MeOH in 0.05% formic acid was
run over 20 min at a flow rate of 0.8 ml/min. The ion peaks and
mass spectra were detected with an Agilent single quadrupole
MSD mass spectrometer with the atmospheric pressure ioniza-
tion-electro spray (API-ES) source in negative and positive ion
modes. The scan range was 100–1000 m/z, and the fragmentor
voltage was 70 V. The VCap was 3000 V for both positive and neg-
ative modes.
a
c
processing statistics for the Os3BGlu6 E178Q complex structures
are presented in Table 1.
Structure refinement
The structures of Os3BGlu6 E178Q with GA4-GE or pNPGlc and
Os3BGlu6 E394D with GA4-GE were solved via rigid body refine-
ment with the wild type Os3BGlu6 structure (PDB: 3GNO) as a
template model. A subset of 5% of the structure factor amplitudes
was reserved for Rfree determination. Manual rebuilding of the
model was performed in Coot [31] and refinement with REFMAC5
[32]. The occupancy of the glucose and glycerol in the active site of
Os3BGlu6 E178Q with GA4-GE was refined by setting the occu-
pancy for the two moieties at values that added to 1.00 or less
and minimizing the positive and negative densities around the li-
gands in the weighted Fo–Fc map, Rfree and B-values. The glucose
bound conformation of E394 (conformation A) was set at the same
occupancy as the glucose, while the sum of the occupancies of the
two E394 conformations was held at 1.00. The overall quality of
each model was evaluated with the program PROCHECK [33].
Additional structure determination and refinement statistics are
presented in Table 1. All the structure figures were drawn with Py-
MOL (Schrödinger, USA).
The transglucosylation product was collected from HPLC sepa-
rations and dried with a vacuum centrifuge. The product was dis-
solved in acetone-d6, and 1H NMR and gCOSY spectra were
collected with a 300 MHz NMR spectrometer (Unity INOVA, Varian,
USA).
Transglucosylation kinetics of the mutants of Os3BGlu6
The pH optima of E178Q and E178A for transglucosylation were
determined in 50 mM MES, pH 5.0–7.5, and 50 mM sodium ace-
tate, pH 4.0–6.0 in 0.5-pH-unit increments. Two micrograms of
E178Q or E178A were incubated with 2 mM GA4-GE and 100 mM
sodium azide, in the different pH buffers at 30 °C for 30 min. The
reactions without sodium azide were used to measure hydrolysis
activities. The reactions were stopped by boiling 1 min and cooled
on ice immediately. The reaction mixes were centrifuged at
6000 rpm for 15 min and separated on LC-MS, as described in the
preceding section. The GA4 released was detected with a diode ar-
ray detector (DAD) at 210 nm, and the amounts were calculated by
comparison of the peak areas to a GA4 standard curve.
Results
The concentrations of donor GA4-GE and acceptor sodium azide
were varied to test their effects on transglucosylation kinetics. The
concentration of sodium azide was varied from 0 to 400 mM, while
the GA4-GE donor was fixed at 2 mM. The reactions were per-
GA4-GE hydrolysis by recombinantly expressed rice GH1 enzymes
GA4-GE, which has previously been reported in rice [24], was
synthesized and used to screen GH1 enzymes for GA4-GE hydroly-

42
Y. Hua et al. / Archives of Biochemistry and Biophysics 537 (2013) 39–48
Table 1
the highest value at pH 4.5, while the activity dropped quickly
Data-collection and structure-refinement statistics.
above pH 5.5 (Supplementary data Fig. S1A). With GA4-GE as sub-
strate, Os3BGlu6 wild type and M251N, E178Q and E178A mutants
all had highest hydrolytic activities at pH 4.5 (Supplementary data
Fig. S1B). The activities of Os3BGlu6 E178Q and E178A dropped
somewhat more slowly at pH above 6.0 compared to Os3BGlu6
wild type and M251N when hydrolyzing GA4-GE, with 50% maxi-
mal activities at approximately 6.5 and 6.0, respectively.
Dataset
Os3BGlu6 E178Q
pNPGlc
Os3BGlu6 E178Q
GA4-GE
PDB code
Beamline
Wavelength (Å)
Space group
Unit-cell parameters (Å)
3WBA
BL13B1
1.00
3WBE
BL13B1
1.00
P212121
a = 57.0
b = 91.1
c = 111.2
30–1.90
1.97–1.90
46370
297256
99.3 (100)
6.5 (6.3)
28.9 (4.5)
6.0 (41.0)
14.8
17.9
478
3917
11 (Glc)
36 (GOL)
534
18.9
16.6
P212121
a = 57.1
b = 91.2
c = 111.4
30–1.97
2.04–1.97
41777
297549
99.9 (100)
7.1 (7.2)
23.3 (6.6)
8.4 (27.6)
14.5
18.0
478
3952
11 (Glc)
36 (GOL)
491
22.6
20.6
Identification of transglucosylation product with azide
Resolution range (Å)
Resolution outer shell(Å)
No. Unique reflections
No. Observed reflections
Completeness (%)
Previously, we found that Os3BGlu6 had negligible transglyco-
sylation activity in reactions with alcohols, such as n-octyl alcohol,
despite its hydrolytic activity toward the corresponding glycosides
(S. Seshadri and J.R. Ketudat Cairns, unpublished). When we tested
Os3BGlu6 for the ability to transfer glucose to azide and GA4 accep-
tors in MES buffer, it hydrolyzed pNPGlc to pNP and glucose, and
no transglucosylation products were observed. In contrast, the
acid/base mutants Os3BGlu6 E178Q and E178A produced transglu-
cosylation products with sodium azide and no significant hydroly-
sis products were detected in overnight reactions (Supplementary
data Fig. S2). However, when the E178Q mutant was used in an at-
tempt to generate GA4-GE by transglycosylation of GA4 acceptor
with pNPGlc donor, no transglycosylation product was detected
(data not shown). When GA4-GE was used as donor and sodium
azide as acceptor for Os3BGlu6 E178Q and E178A, a prominent
transglucosylation product and a smaller amount of glucose were
observed (Supplementary data Fig. S3).
Average redundancy per shell
I/
r (I)
R
(%)
(merge)
Rfactor (%)
Rfree (%)
No. of residues in protein
No. Protein atoms
No. Ligand atoms
No. Other hetero atoms
No. waters
Mean B-factor
Protein
Ligand
Other hetero atoms
Waters
r.m.s. bond deviations (length)
r.m.s angle deviations (degrees)
Ramachandran plot
Residues in most favorable
regions (%)
Residues in additional allowed
regions (%)
Residues in generously allowed
regions (%)
12.1
35.3 (GOL)
34.3
0.013
1.290
11.1
44.2 (GOL)
36.8
0.013
1.264
The transglucosylation product of E178A with GA4-GE was
identified with LC-MS (Fig. 1). The peak of newly formed product
at retention time 3.69 min had the molecular mass consistent with
88.9
10.7
0.5
88.6
10.9
0.5
the 205.17 a.m.u mass of b-D-glucopyranosyl azide (1-azido-b-D-
glucoside) expected from the retaining mechanism of the b-gluco-
sidase [34]. Since formic acid (HCO2H) was used in the mobile
phase of LC-MS, its adduct ion [M+HCO2]ꢁ was detected as the base
peak along with [M–H]ꢁ and [M+Cl]ꢁ peaks (Supplementary data
Residues in disallowed regions
(%)
0.0
0.0
Fig. S4). The structure of the b-D-glucopyranosyl azide was also
confirmed from its 1H and gCOSY NMR spectra (data not shown).
The 1H spectrum peaks were assigned as: d4.53, H1, d,
sis activity. Five rice GH1 enzymes that have been expressed in
E. coli were tested for the hydrolysis of pNPGlc and GA4-GE. As
shown in Table 2, Os3BGlu6 was found to have the highest hydro-
lysis activity to GA4-GE among these enzymes. Although Os9B-
Glu31 had a higher ratio of activity toward GA4-GE compared to
pNPGlc (0.27 vs. 0.07 for Os3BGlu6), it is primarily a transglycosi-
dase [29] and has low activity toward both substrates.
J2,1 = 9.0 Hz; d3.85, H6, dd, J6,6 = 11.1 Hz; d3.678, H60, dd,
0
0
0
J5,6 = 4.8 Hz, J6,6 = 11.1 Hz; d3.364–3.458, H3, H4, H5, m; d3.192,
H2, t, J1,2 = J3,2 = 9.0 Hz, which matched the published 1H NMR data
for b-D-glucopyranosyl azide [35]. The coupling constant between
H1 and H2 of 9.0 Hz confirmed the b-
the ‘‘b’’ configuration.
D-glucopyranosyl azide had
Hydrolysis activities of Os3BGlu6 and its mutants
Transglucosylation kinetics of the mutants of Os3BGlu6
Mutation of the putative catalytic acid/base and nucleophile
residues and aglycone-binding residue M251 had differential ef-
fects on pNPGlc and GA4-GE hydrolysis. The mutation of the Os3B-
Glu6 glutamate residue 178 to alanine (E178A) and glutamine
(E178Q) and glutamate residue 394 to aspartate (E394D) and glu-
tamine (E394Q) reduced the relative hydrolytic activity toward
pNPGlc to <1%, while the M251N mutant reduced this activity to
52% (Table 3). The relative activity of M251N was 88.9% for GA4-
GE, and the catalytic efficiency (kcat/Km) followed the same pattern
with decreases from 6.2 to 2.6 mMꢁ1 sꢁ1 for pNPGlc, and from 0.13
to 0.08 mMꢁ1 sꢁ1 for GA4-GE compared to its wild type. For GA4-
GE, E178A had 22.2% and E178Q 12.5% of the relative hydrolytic
activity of wild type Os3BGlu6 with both Km and kcat values lower
than wild type. In contrast, the E394Q and E394D mutants had <2%
wild type relative hydrolytic activity with GA4-GE, similar to that
with pNPGlc.
The activity versus pH profiles for transglucosylation of azide
with GA4-GE donor were determined in sodium acetate and MES
buffers (Supplementary data Fig. S5). For the Os3BGlu6 E178Q mu-
tant, its optimum pH range was 5.0 to 6.0 in both buffers, slightly
higher than its optimum pH for hydrolysis of GA4-GE. The E178A
mutant also had highest transglucosylation activity at pH 5.0 in
the MES buffer, but was high at pH 4.0 in the sodium acetate. In
the sodium acetate buffer, the release of pNP from pNPGlc by
E178A was much higher than that in MES buffer, even without
azide in the system.
Fig. 2 shows that when the concentration of donor GA4-GE was
fixed at 2 mM, the turnover rates, V0/E0, of GA4 for Os3BGlu6 E178A
slowly increased with increasing concentrations of sodium azide
and reached its maximum at 400 mM sodium azide. However, for
Os3BGlu6 E178Q, the turnover rates of GA4 decreased at higher
azide concentrations after reaching its maximum at 100 mM so-
dium azide. Since transglucosylation products were the main prod-
ucts and little glucose was evident when the azide acceptor was
The Os3BGlu6 wild type and its M251N mutant were found to
have high pNPGlc hydrolysis activity between pH 4.0 to 5.0, with

Y. Hua et al. / Archives of Biochemistry and Biophysics 537 (2013) 39–48
43
Table 2
loop of residues 331 to 337, which was poorly defined by the elec-
tron density, as was the N-terminus (similar to the previously re-
ported wild type structure). Soaking of Os3BGlu6 E394D crystals
with GA4-GE resulted in electron density in both the glycone and
aglycone subsites, but the occupancy was too low and the ligand
density too poorly defined to build a reliable GA4-GE complex
structure.
GA4-GE hydrolysis by recombinantly expressed rice GH1 enzymes.
Enzyme
Phylogenetic
clustera
Activity
toward GA4-
Activity
toward pNPGlc activity
Ratio of
GE (lmol Glc
(
lmol pNP
toward
GA4-GE/
pNPGlc
released/min/
mg)
released/min/
mg)
Os3BGlu6
Os3BGlu7
(BGlu1)
At/Os1
At/Os4
0.185
0.02
2.6
4.0
0.07
0.005
In the covalent intermediate structure, the anomeric carbon of
glucose is covalently bonded with Oe1 of E394 at a distance of
1.3 Å (Figs. 3A and 4A). The noncovalent interactions between
the surrounding amino acids and the hydroxyls of the glucose ring,
which took a 4C1 chair conformation in the ꢁ1 subsite, included ten
hydrogen bonds. The O2 atom of glucose forms hydrogen bonds
Os4BGlu12 At/Os7
Os4BGlu18 At/Os5
Os9BGlu31 At/Os6
0.035
N.D.
130
0.003
–
0.27
0.94
0.02b
0.075b
N.D. means not detectable.
Phylogenetic clusters from the eight sequence-based GH1 phylogenetic clusters
shared by rice and Arabidopsis, as identified by Opassiri et al. [28].
a
with H132N
(2.99 Å), while O3 of glucose hydrogen bonds with Q31 O
(2.66 Å) and W452N
bonded to Q31N 2 (2.91 Å) and E451 O
cose forms a hydrogen bond with Y321 (2.70 Å), while O6 forms a
hydrogen bond with O 2 of the same residue (2.60 Å). A water
molecule (Water52) is also hydrogen bonded to O2 of glucose
(2.80 Å) and E394 O 1 (2.98 Å) (Fig 4A). A glycerol molecule was
e2
(3.30 Å), N177
Oe
1
(3.01 Å) and Q178N
e
2
e
1
b
Activity is primarily transglycosylation, rather than hydrolysis [29].
e1 (2.85 Å). The glucosyl O4 is hydrogen
e
e
1 (2.61 Å), and O5 of glu-
present compared to when it was not, the parameters were calcu-
lated from the released GA4 without subtracting hydrolysis. The
Os3BGlu6 E179Q and E179A mutants had similar kinetic parame-
ters for sodium azide in MES buffer, in the presence of 2 mM
GA4-GE with apparent Km of 3.71 0.47 mM, kcat of
0.48 0.01 sꢁ1 and kcat/Km of 0.13 mMꢁ1 sꢁ1 for E178Q, and Km of
e
e
observed in the active site of this Os3BGlu6 E178Q structure in
two conformations in the +1 subsite.
3.1 0.5 mM, kcat of 0.204 0.006 sꢁl, and kcat/Km of 0.066 mMꢁ1
-
sꢁ1 for E178A.
Discussion
The effects of the concentration of sodium azide on the trans-
glucosylation kinetics were further studied by measuring kcat and
Km for GA4-GE at 50, 100, 200 and 400 mM sodium azide (Table 4).
For Os3BGlu6 E178Q, the Km for GA4-GE was >18-fold higher and
kcat at least 27-fold higher than that without sodium azide at all
concentrations tested and both the Km and kcat increased with
azide concentration, while the kcat/Km of GA4-GE was highest at
0.51 mMꢁ1 sꢁ1 in 100 mM sodium azide and decreased to 0.41
and 0.29 when the concentration of sodium azide was increased
to 200 and 400 mM, respectively. For Os3BGlu6 E178A, the Km
was 3-fold higher and kcat 15-fold higher with 200 mM sodium
azide compared to without sodium azide, while the kcat/Km of
GA4-GE was the same at 200 mM and 400 mM sodium azide.
Hydrolysis of GA4-GE by Os3BGlu6 and its mutants
Thirty-four GH1 genes encoding potentially active rice b-gluco-
sidases have been reported, and these were broken into eight se-
quence-based phylogenetic clusters that were shared between
rice and Arabidopsis thaliana [28]. The five rice GH1 enzymes tested
each represented a different phylogenetic cluster, from which we
have been able to express active enzymes. Four of the tested en-
zymes had the ability to free GA4 from GA4-GE, and could poten-
tially contribute to this release in planta. Although it also
hydrolyzes b-1,2- and b-1,3-linked gluco-disaccharides and octyl-
b-D-glucoside with high activity and may play other roles in planta
[26], Os3BGlu6 had the highest activity toward GA4-GE and was
chosen to characterize the hydrolysis of the glucosyl ester.
Although much characterization has been done of the kinetics
of hydrolysis of alkyl and aryl glycosides and oligosaccharides by
Structure of Os3BGlu6E178Q soaked with either GA4-GE or pNPGlc
When the Os3BGlu6 E178Q (acid/base) mutant was soaked with
GA4-GE, electron density for a glucosyl residue covalently bound to
the catalytic nucleophile was seen, rather than a GA4-GE molecule.
However, when the glucosyl residue was built into the density, a
patch of positive density protruded in the axial position from glu-
cose carbon 4 and another positive density patch corresponded to
the position of the E394 sidechain in the apo Os3BGlu6 structure
(PDB: 3GNO). When the two sidechain positions and glucosyl moi-
ety were set at 50% occupancy, a positive density defining a glyc-
erol molecule was observed at positions overlapping the glucosyl
C2, O2, C3, O3 and C4, but pointing away from O4. Thus, the glu-
cose was present with partial occupancy, with a glycerol molecule
from the solvent sharing the same positions as glucose carbons 2, 3
and 4 (PDB: 3WBE; Figs. 3B and 4B). The catalytic nucleophile was
found in both the position previously seen for the covalent com-
plex of Os3BGlu6 with 2-deoxy-2-fluoroglucoside (G2F) and that
found in the apo enzyme (Figs. 3B, 4B, and 5C) [26]. Another crystal
b-D-glucosidases and their catalytic acid/base and nucleophile mu-
tants [34–37], little description of the kinetics of glucosyl ester
hydrolysis is available in the literature. Kiso et al. [38] compared
the activities of 3 bacterial, 2 fungal and almond b-glucosidases
for hydrolysis of 1-O-(p-hydroxybenzyoyl) b-D-glucose (pHBGlc),
and found that Caldocellum saccharolyticum b-glucosidase had the
highest activity of the commercial enzymes tested with 21% activ-
ity for pHBGlc relative to pNPGlc. Almond was reported to have ap-
prox. 5% relative activity for pHBGlc, somewhat less than the 7% of
Os3BGlu6 for GA4-GE compared to pNPGlc. In contrast to Caldocel-
lum saccharolyticum and almond b-glucosidase, which had similar
Vmax for pNPGlc and pHBA but nearly 10-fold and 15-fold higher
Km for pHBA, respectively [39], Os3BGlu6 had similar Km values
for pNPGlc and GA4-GE, but a 26-fold lower apparent kcat for
GA4-GE relative to pNPGlc. Although the large aglycone of the
GA4-GE is not strictly comparable to that of pNPGlc, it shows that,
in the case of Os3BGlu6, the rate of the hydrolysis and not the
binding step is the distinguishing feature of the ester compare to
the glucoside.
soaked in 2 mM pNPGlc also showed a covalently bound a-D-gluco-
syl residue, but with high occupancy (PDB: 3WBA; Figs. 3A and
4A). Since the structure appeared the same, but the ligand density
was clearer in the pNPGlc structure (Fig. 4A), this structure best
represents the covalent complex of the Os3BGlu6 E178Q mutant
with an unmodified glucosyl residue. In both structures, the pro-
tein structure was well defined from G11 to T488, except for the
It might be expected that the relatively low pKa of the leaving
group would make the glucosyl ester relatively insensitive to the
presence of a catalytic acid for the initial glycosylation step (the
GA4 pKa is 4.3, compared to 7.2 for pNPGlc), similar to 2,4-dinitro-

44
Y. Hua et al. / Archives of Biochemistry and Biophysics 537 (2013) 39–48
Table 3
Hydrolysis activities of Os3BGlu6 and its mutants.
Enzyme
pNPGlc
GA4-GE
Relative activity (%)
Km (mM)
kcat (sꢁ1
)
kcat/Km (mMꢁ1 sꢁ1
)
Relative activity (%)
Km (mM)
kcat (sꢁ1
)
kcat/Km (mMꢁ1 sꢁ1
)
Os3BGlu6
100
51.6
0.3
0.7
0.9
6.3 0.4a
6.5 0.7
n.m.
n.m.
n.m.
38.9 0.9a
15.4 0.9
n.m.
n.m.
n.m.
6.2a
2.6
n.m.
n.m.
n.m.
n.m.
100
88.9
12.5
22.2
1.4
5.8 0.6
14.6 2.0
0.09 0.01
3.7 0.4
n.m.
0.75 0.04
1.2 0.1
0.03 0.001
0.09 0.005
n.m.
0.13
0.08
0.33
0.03
n.m.
n.m.
Os3BGlu6-M251N
Os3BGlu6-E178Q
Os3BGlu6-E178A
Os3BGlu6- E394D
Os3BGlu6-E394Q
0.8
n.m.
n.m.
0.6
n.m.
n.m.
n.m.: not measured.
For the relative activity with pNPGlc, 0.5
(final volume 50 l) at 30 °C for 10 min. For the relative activity with GA4-GE, 1.0
incubated with 1.7 mM GA4-GE in NaOAc, pH 4.5 (final volume was 50 l), at 30 °C for 20 min. The relative activities were calculated based on A405 and the amount of the
lg of Os3BGlu6 wild type or M251 N, or 3
lg of E178A, E178Q, E394D or E394Q was incubating with 4 mM pNPGlc in NaOAc, pH 4.5
l
l
g of Os3BGlu6 wild type or M251 N, or 3 g of E178A, E178Q, E394D or E394Q was
l
l
enzymes. For determination of the kinetic parameters, 50 mM NaOAc buffer was used for Os3BGlu6 wild type and M251 N, for both substrates, but 50 mM MES buffer was
selected for the E178Q and E178A mutants.
a
Data from Seshadri et al. (2009).
phenyl glycosides [34,41], but other influences of having the ester
bond in apposition to the glycosidic bond are not so clear. The
Os3BGlu6 wild type had an optimum pH for hydrolysis of both
pNPGlc and GA4-GE at pH 4.5, indicating that the low pKa of the
leaving group carboxylate did not affect the pH optima. Although
the Os3BGlu6 E178Q and E178A mutants had low activities toward
pNPGlc, their GA4-GE hydrolysis activities were significant and
dropped slowly above pH 6.0, compared to wild type Os3BGlu6,
consistent with their designations as catalytic acid/base mutants.
The pH dependence of enzymatic reactions is generally considered
to reflect ionizations of acid/base groups involved in catalysis
[39,40]. The Os3BGlu6 E178Q and E178A mutants lack the ioniz-
able group to donate a proton in the glycosylation step or extract
a proton in the deglycosylation step, resulting in lower pH depen-
dence at high pH.
In contrast to the Agrobacterium and rice Os3BGlu7 b-glucosi-
dases [37,41], the acid/base glutamate to alanine (E178A) mutant
was more active than the acid/base to glutamine mutant
(E178Q). This was evidently due to the buildup of an intermediate
on the enzyme, since both the Km and kcat dropped dramatically in
the E178Q mutant. The X-ray crystallographic structure (Fig. 3)
indicated this intermediate was the covalent intermediate formed
in the glycosylation half reaction (Fig. 5). The glutamine sidechain
can provide a hydrogen bond to stabilize the leaving group in gly-
cosylation and acceptor in deglycosylation of the enzyme, giving it
higher activity [42], but it evidently could not facilitate water as an
acceptor in Os3BGlu6. Without E178, the low pKa of GA4 allows it
to be released in the glycosylation step without general acid-assis-
tance more effectively than pNPGlc, for which glycosylation ap-
peared to be the rate-limiting step. The lack of deprotonation of
the water molecule in the deglycosylation step makes Os3BGlu6
E178A and Os3BGlu6 E78Q relatively effective at transglycosyla-
tion with GA4-GE donor compared to their wild type.
When the structure of wild type Os3BGlu6 was resolved by X-
ray crystallography, the M251 residue was suggested to block
binding of straight cellooligosaccharides in the active site cleft
[26]. The mutation of M251 to asparagine resulted in 9 to 24-fold
increases hydrolysis efficiency for oligosaccharides in Os3BGlu6
M251N compared to wild type [43]. So, it was of interest to see
whether this mutation would affect the hydrolysis of less polar
substrates like pNPGlc and GA4-GE as well. The catalytic efficien-
cies (kcat/Km) for M251N were reduced 2.4-fold for pNPGlc, and
1.6-fold for GA4-GE compared to wild type. The Km to pNPGlc for
both Os3BGlu6 and Os3BGlu6 M251N is almost the same, and
the smaller kcat of Os3BGlu6 M251N is the main cause for its lower
k
cat/Km compared to the wild type. In contrast, for GA4-GE, the
higher Km for Os3BGlu6 M251N is the main reason for its lower
cat/Km compared to the wild type, suggesting that the larger agly-
k
cone of GA4-GE may interact with residue M251, which may stabi-
lize binding of large hydrophobic aglycones.
Fig. 1. The LC-MS chromatogram of the transglucosylation reaction of Os3BGlu6 E178A with GA4-GE as donor and sodium azide as acceptor. The reaction components were
separated by C18 reverse phase HPLC, as described in the methods. The detector was an Agilent MSD single quadrupole mass spectrometer with API-ES source in negative ion
mode, with a scan range of 50–800 m/z.

Y. Hua et al. / Archives of Biochemistry and Biophysics 537 (2013) 39–48
45
30.0
25.0
20.0
15.0
10.0
5.0
E178A
E178Q
0.0
0
50
100
150
200
250
300
350
400
Na-azide (mM)
Fig. 2. Rates of transglucosylation of azide acceptor with GA4-GE donor by the Os3BGlu6 E178A and E178Q acid/base mutants. The concentrations of sodium azide acceptor
were varied from 0 to 400 mM, while donor GA4-GE was fixed at 2 mM. The reactions were performed in 50 mM MES, pH 5, at 30 °C for 20 min. The turnover rates V0/E0 of
GA4 (lmole of GA4 release per minute per lmole of enzyme) were calculated based on the area of GA4 HPLC peaks. The total area of GA4 was used without subtracting the
hydrolysis reaction.
GE to produce b-D-glucosyl azide, as shown schematically in
Table 4
Transglucosylation kinetics of Os3BGlu6 E178Q and Os3BGlu6 E178A for GA4-GE
donor at various fixed concentrations of sodium azide acceptor.
Fig. 5. In this case, Os3BGlu6 E178Q had higher activity than
E178A. For the Os3BGlu6 E178A, the GA4 releasing activity in so-
dium acetate was much higher than that in MES buffer, even with-
out azide in the system, indicating that acetate may also act as a
nucleophile and/or substitute acid/base to rescue the activity, as
was seen in the work of Wang et al. [35].
The transglucosylation turnover rate, V0/E0, of GA4 increased
very significantly as the concentration of azide was increased.
The rate increase with increasing concentration of azide was also
observed by Wang et al. [35]. They concluded that addition of azide
as a competitive nucleophile increased kcat values 100–300 fold for
substrates for which the rate-limiting step is deglycosylation, such
as GA4-GE with Os3BGlu6 E178Q and Os3BGlu6 E178A. As the con-
centration of azide is increased, the rate of deglycosylation in-
creases, until glycosylation becomes rate determining. The build-
up at the E178Q covalent intermediate was released by reaction
with the azide, which was apparently facilitated by interaction of
Q178 with the azide. Although the Km values for azide for the
Enzyme
Sodium azide
(mM)
Km (mM)
kcat (sꢁ1
)
kcat/Km
(mMꢁ1sꢁ1
)
Os3BGlu6
E178Q
0
50
0.09 0.01 0.03 0.001 0.33
1.64 0.15 0.81 0.04
1.69 0.13 0.86 0.04
2.07 0.31 0.84 0.08
3.97 0.65 1.16 0.14
3.7 0.4
11.6
0.49
0.51
0.41
0.29
100
200
400
0
200
400
Os3BGlu6
E178A
0.09 0.005 0.03
1.39
1.56
0.12
0.12
13.36
The transglucosylation rates were calculated from the released GA4 without sub-
tracting hydrolysis product.
Transglucosylation activities of Os3BGlu6E178Q and Os3BGlu6E178A
When azide was included in the reaction, both Os3BGlu6 E178A
and E178Q could catalyze transglycosylation reactions from GA4-
Fig. 3. The electron density map (Omit Fobs ꢁ Fcalc contoured at 3
r) of glycosyl-enzyme intermediate of the Os3BGlu6 E178Q with pNPGlc (A) and GA4-GE (B). (A) The glucose
of Os3BGlu6 E178Q soaked with pNPGlc are represented by balls and sticks with carbon in yellow for glucose. The nucleophile residue (E394) is represented by sticks with
carbon in gray, oxygen in red and nitrogen in dark blue. (B) The mixed structures of apo Os3BGlu6 E178Q and Os3BGlu6 E178Q covalently bound to Glc from GA4-GE was
modeled with two conformations of E394 and glucose and glycerol occupying the same position, by refining the occupancy to 0.5 for glucose and glycerol and each of the
E394 conformations. The glucose and glycerol are shown in ball and stick representation with carbon in cyan for glucose and in green for glycerol. The nucleophile residue
(E394) is represented as sticks with carbon in gray, oxygen in red and nitrogen in dark blue. (C) A top view of the glucosyl ring and the overlapping glycerol electron density
for Os3BGlu6 E178Q with GA4-GE shown in B to clarity the different positions of the glucosyl 4-OH and glycerol 3-OH. (For interpretation of the references to colour in this
figure legend, the reader is referred to the web version of this article.)

46
Y. Hua et al. / Archives of Biochemistry and Biophysics 537 (2013) 39–48
Fig. 4. Binding of glucose in the active site of Os3BGlu6 E178. A and B, Stereo view of protein–ligand interaction in the active site of the Os3BGlu6 E178Q soaked with pNPGlc
and with GA4-GE, respectively. The amino acids surrounding the ꢁ1 subsite are represented as sticks with carbon in gray, nitrogen in dark blue and oxygen in red. The
covalent glucosyl moiety is represented as balls and sticks with carbon in yellow for Os3BGlu6 E178Q soaked with pNPGlc and in cyan for Os3BGlu6 E178Q soaked with GA4-
GE, and oxygen in red. The two glycerols in both complex structures are represented as balls and sticks with carbon in green and oxygen in red. Hydrogen bonding
interactions of glucose and glycerol with amino acid residues are shown as black dotted lines.
two mutants were similar, the kcat/Km values for sodium azide as
acceptor with 2 mM GA4-GE donor of 0.13 mMꢁ1 sꢁ1 for Os3BGlu6
E178Q and 0.066 mMꢁ1 sꢁ1 for Os3BGlu6 E178A indicate that the
polar glutamine at residue 178 in Os3BGlu6 E178Q supports trans-
glycosylation by azide better than the nonpolar alanine residue in
Os3BGlu6 E178A. A stronger interaction with the azide in the
Os3BGlu6 E178Q mutant is also supported by the fact that it is
inhibited by high concentrations of azide, while the E178A mutant
is not.
Structures of Os3BGlu6 covalent intermediate complexes
Although we intended to obtain a noncovalent GA4-GE complex
with an Os3BGlu6 mutant, the high reactivity of this substrate re-
sulted in poor density in the E394D nucleophile mutant active site
and the covalent glucosyl-enzyme intermediate in the E178Q acid/
base mutant. Fig. 6 shows that the positions and interactions made
by amino acid residues surrounding the native glucosyl moiety in
the active site of Os3BGlu6 E178Q with glucose from GA4-GE
Fig. 5. Schematic of the hydrolysis and transglycosylation reactions catalyzed by Os3BGlu6 and its acid/base mutants with GA4-GE. The first step, glycosylation, is allowed to
proceed without acidic assistance to release the GA4, due to its low pKa. The covalent intermediate, seen in the crystal structures, builds up in the acid/base mutants E178A
and E178Q, but the presence of a small nucleophile (Nuꢁ) such as azide promotes deglycosylation of the mutant enzymes by transglycosylation. In contrast, the wild type
Os3BGlu6 only appears to catalyze hydrolysis.

Y. Hua et al. / Archives of Biochemistry and Biophysics 537 (2013) 39–48
47
Fig. 6. Comparison of the active site region of Os3BGlu6 bound with G2F (A) and those of the structures of Os3BGlu6 E178Q soaked with pNPGlc (B) and Os3BGlu6 E178Q
soaked with GA4-GE (C). The G2F and glucosyl moieties are represented as balls and sticks with carbon in pink for G2F, in yellow for Os3BGlu6 E178Q soaked with pNPGlc and
cyan for Os3BGlu6 E178Q soaked with GA4-GE, oxygen in red and fluorine in pale cyan. The amino acid residues surrounding the ligands are represented by sticks with
carbons in gray, nitrogens in dark blue and oxygens in red for all three structures. (D) Superpositioin of the three structures in (A–C), showing the nucleophile residues
covalently bound to the glucosyl moiety at the ꢁ1 subsite of the Os3BGlu6E178Q structures soaked with pNPGlc (yellow carbons) and GA4-GE_(cyan carbons) and to the G2F
moiety in Os3BGlu6 (3GNP, pink carbons). The unbound position of the nucleophile is represented by the alternative conformation of Os3BGlu6 E178Q soaked with GA4-GE.
The molecules are represented as in (A–C), except for the protein carbon colors.
(PDB: 3WBE) or pNPGlc (PDB: WBA) were similar to those interact-
ing with the 2-deoxyl-2-fluroglucopyranoside residue (G2F) in the
structure of wild type Os3BGlu6 with G2F (3GNR). In the structure
of Os3BGlu6/G2F complex [26], the density of the 2-deoxy-2-flu-
oroglucosyl residue matched a relaxed 4C1 chair conformation, as
did the glucosyl moiety in the Os3BGlu6 E178Q covalent interme-
diate structure (Fig 6D). The distance between the anomeric carbon
of glucose and the Glu394 nucleophile residue of Os3BGlu6 E178Q
was also similar to that in the G2F covalent complex (1.3 Å). A
water molecule was observed between the 2-OH of glucose or fluo-
rine atom of G2F and N177 and E394 in the Os3BGlu6 E178Q/Glc
and Os3BGlu6/G2F structures. A water molecule in this position
has been shown to be critical for the hydrolysis and glycosynthase
reactions of Os3BGlu7 [43]. The hydrogen bonding pattern be-
tween glucosyl residues at the ꢁ1 subsite of the Os3BGlu6
E178Q/Glc covalent complex structure is similar when compared
to that of the Os3BGlu6/G2F complex (Fig. 6A, B and C). The aver-
age distances between the 2-OH of the glucose and 2-fluorine of
G2F moieties and polar atoms of the surrounding amino acid resi-
dues are similar (Fig 6A–D).
lation [50]. The human cytosolic b-glucosidase acid/base mutant
glucose complex (2ZOX) is the only previously reported GH1 struc-
ture with the nucleophile covalently bound to native glucose [51].
The hydrogen bonding pattern between G2F molecule and the en-
zyme at the ꢁ1 subsite is conserved in other reported GH1 cova-
lent intermediates, but this pattern lacks the hydrogen bonds for
which the 2-OH is the proton donor seen in human cytosolic b-glu-
cosidase and Os3BGlu6 E178Q covalent glucoside complex struc-
tures. In general, the reports of G2F complexes do not report
hydrogen bonds to the 2-F group, but the conservation of the water
in proximity to the fluorine in these complexes supports the idea
that the fluorine may accept a hydrogen bond from the water
and other nearby hydrogen donors. On the other hand, the loss
of possible hydrogen bonds from the 2-OH to Od of N177 and
Oe2 of E394 in Os3BGlu6, which are seen in the native glucosyl-en-
zyme intermediates (Fig. 6), may result in a further loss of stabil-
ization of the transition state in the 2-F-glucoside reaction [51].
Conclusion
Many structures of GH1 glycosyl-enzyme intermediates with a
2-deoxy-2-fluoroglucosyl (G2F) moiety bound to the nucleophilic
residue have been reported, including one for Os3BGlu6 (PDB code
3GNR) [26,44–49]. These GH1 glycosyl-enzyme intermediates used
2,4-dinitrophenyl 2-deoxy-2-fluoroglucoside, with which the sub-
stitution of an electronegative fluorine atom for a hydroxyl group
adjacent to the reaction center at C-2 destabilizes the transition
states and decreases the rates of both glycosylation and deglycosy-
This work identified Os3BGlu6 as a b-glucosidase with rela-
tively high activity to hydrolyze GA4-GE. Comparison of the activ-
ity of Os3BGlu6 and its acid/base mutants, E178Q and E178A, for
hydrolysis of pNPGlc and GA4-GE suggested that hydrolysis of es-
ters and glycosides is similar and the low pKa of the carboxyl leav-
ing group is what differentiates it, although the rate of the wild
type enzyme is much slower with the ester. Despite the
observation of little or no transglucosylation activity for wild type

48
Y. Hua et al. / Archives of Biochemistry and Biophysics 537 (2013) 39–48
[19] T. Yoshihara, M. Amanuma, T. Tsutsumi, Y. Okumura, H. Matsuura, A. Ichihara,
Plant Cell Physiol. 37 (1996) 586–590.
[20] S. Wakuta, S. Hamada, H. Ito, R. Imai, H. Mori, H. Matsuura, K. Nabeta, H.
Matsui, J. Appl. Glycosci. 58 (2011) 67–70.
Os3BGlu6, its acid/base mutants were still rescued by azide trans-
glycosylation to form b- -glucosyl azide. The covalent intermedi-
ate of Os3BGlu6 E178Q with unmodified -glucoside was
D
a-D
[21] P.J. Davies, Plant Hormones: Physiology, Biochemistry and Molecular Biology,
Kluwer Academic Publishes, 1995.
[22] G. Schneider, in: A. Crazier (Ed.), The Biochemistry and Physiology of
Gibberellins, vol. 1, Praeger Publishers, New York, 1983, pp. 389–456.
[23] G. Schneider, E. Jensen, C.R. Spray, B.O. Phinney, Proc. Natl. Acad. Sci. USA 89
(1992) 8045–8048.
observed with either pNPGlc or GA4-GE donor substrate, support-
ing the covalent double displacement mechanism for this enzyme
with either glycoside or glycosyl ester substrates.
Acknowledgments
[24] M. Koshioka, E. Minami, H. Saka, R.P. Pharis, L.N. Mander, Gibberellins,
Springer, New York, 1991. 264–272.
The authors are grateful to Assist. Prof. Dr. Thanaporn Manyum
for providing her laboratory for initial synthesis work and Ms.
Supaporn Baiya for providing Os4BGlu18 b-glucosidase. Suranaree
University of Technology (SUT) and the Center for Scientific and
Technological Equipment (CSTE) are thanked for supporting YH
in her Ph.D. studies. The research funding was provided by SUT
and the National Research University Project of the Commission
on Higher Education of Thailand, along with the Thailand Research
Fund (Grant BRG5380017).
[25] K. Hiraga, T. Yokota, N. Takahashi, Phytochemistry 13 (1974) 2371–2376.
[26] S. Seshadri, T. Akiyama, R. Opassiri, B. Kuaprasert, J.R. Ketudat Cairns, Plant
Physiol. 151 (1) (2009) 47–58.
[27] R. Opassiri, J.R. Ketudat Cairns, T. Akiyama, O. Wara-aswapati, J. Svasti, A. Esen,
Plant Sci. 165 (2003) 627–638.
[28] R. Opassiri, B. Pomthong, T. Onkoksoong, T. Akiyama, A. Esen, J.R. Ketudat
Cairns, BMC Plant Biol. 6 (2006) 33.
[29] S. Luang, J.-I. Cho, B. Mahong, R. Opassiri, T. Akiyama, K. Phasai, J. Komvongsa,
N. Sasaki, Y. Hua, Y. Matsuba, Y. Ozeki, J.-S. Jeon, J.R. Ketudat Cairns, J. Biol.
Chem. 288 (2013) 10111–10123.
[30] Z. Otwinowski, W. Minor, Meth. Enzymol. 276, Academic Press, New York,
(1997) 307–326.
[31] P. Emsley, K. Cowtan, Acta Crystallogr. D 60 (2004) 2126–2132.
[32] G.N. Murshudov, A.A. Vagin, E.J. Dodson, Acta Crystallogr. D 53 (1997) 240–
255.
Appendix A. Supplementary data
[33] R.A. Laskowski, M.W. MacArthur, D.S. Moss, J.M. Thornton, Appl. Cryst. 26
(1993) 283–291.
[34] Q. Wang, D. Trimbur, R. Graham, R.A.J. Warren, S.G. Withers, Biochemistry 34
(1995) 14554–14562.
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.abb.2013.06.005.
[35] Q. Wang, R.W. Graham, D. Trimbur, R.A.J. Warren, S.G. Withers, J. Am. Chem.
Soc. 116 (1994) 11594–11595.
References
[36] L. Mackenzie, G. Sulzenbacher, C. Divne, T.A. Jones, H.F. Woldike, M. Schulein,
S.G. Withers, G. Davies, Biochem. J. 335 (1998) 409–416.
[37] W. Chuenchor, S. Pengthaisong, R.C. Robinson, J. Yuvaniyama, J. Svasti, J.R.
Ketudat Cairns, J. Struct. Biol. 173 (2011) 169–179.
[38] T. Kiso, S. Kitahata, K. Okamoto, S. Miyoshi, H. Nakano, J. Biosci. Bioeng. 90
(2000) 614–618.
[39] J.B. Kempton, S.G. Withers, Biochemistry 31 (1992) 9961–9969.
[40] L.P. McIntosh, G. Hand, P.E. Johnson, M. Joshi, M. Korner, L.A. Plesniak, L. Ziser,
W.W. Wakarchuk, S.G. Withers, Biochemistry 35 (1996) 9958–9966.
[41] J. Müllegger, M. Jahn, H.M. Chen, R.A. Warren, S.G. Withers, Protein Eng. Des.
Sel. 18 (2005) 33–40.
[42] S. Sansenya, J. Maneesan, Carbohydr. Res. 351 (2012) 130–133.
[43] J. Wang, S. Pengthaisong, J.R. Ketudat Cairns, Y. Liu, Biochim. Biophys. Acta.
1834 (2013) 536–545.
[1] T. Fowler, in: A. Esen (Ed.), Biochemistry and Molecular Biology, American
Chemical Society, Washington, DC, 1993, pp. 56–65. ACS Symposium Series
533.
[2] J.E. Poulton, Plant Physiol. 94 (1990) 401–405.
[3] L. Duroux, F.M. Delmotte, J.M. Lancelin, G. Keravis, C. Jay-Alleand, Biochem. J.
333 (1998) 275–283.
[4] B. Brzobohaty´, I. Moore, P. Kristoffersen, L. Bako, N. Campos, J. Schell, K. Palme,
Science 262 (1993) 1051–1054.
[5] A. Falk, L. Rask, Plant Physiol. 108 (1995) 1369–1377.
[6] D.P. Dharmawardhana, B.E. Ellis, J.E. Carlson, Plant Physiol. 107 (1995) 331–
339.
[7] M. Reuveni, Z. Sagi, D. Evnor, A. Hetzroni, Plant Sci. 147 (1999) 19–24.
[8] L. Barleben, X. Ma, J. Koepke, G. Peng, H. Michel, J. Stöckigt, Biochim. Biophys.
Acta. 1747 (2005) 89–92.
[9] B. Henrissat, Biochem. J. 280 (1991) 309–316.
[10] B.L. Cantarel, P.M. Coutinho, C. Rancurel, T. Bernard, V. Lombard, B. Henrissat,
Nucleic. Acids Res. 37 (2009) D233–D238.
[44] W. Burmeister, S. Cottaz, H. Driguez, R. Iori, S. Palmieri, B. Henrissat, Structure
5 (1997) 663–675.
[45] P. Isorna, J. Polaina, L. Latorre-García, F.J. Cañada, B. González, J. Sanz-Aparicio,
J. Mol. Biol. 371 (2007) 1204–1218.
[46] T.M. Gloster, S. Roberts, V.M.A. Ducros, G. Perugino, M. Rossi, R. Hoos, M.
Moracci, A. Vasella, G.J. Davies, Biochemistry 43 (2004) 6101–6109.
[47] D.L. Zechel, S.P. Reid, D. Stoll, O. Nashiru, R.A. Warren, S.G. Withers,
Biochemistry 42 (2003) 7195–7204.
[48] W. Chuenchor, S. Pengthaisong, R.C. Robinson, J. Yuvaniyama, W. Oonanant,
D.R. Bevan, A. Esen, C.J. Chen, R. Opassiri, J. Svasti, J.R. Ketudat Cairns, J. Mol.
Biol. 377 (2008) 1200–1215.
[49] S. Sansenya, R. Opassiri, B. Kuaprasert, C.J. Chen, J.R. Ketudat Cairns, Arch.
Biochem. Biophys. 510 (2011) 62–72.
[50] J. Noguchi, Y. Hayashi, Y. Baba, N. Okino, M. Kimura, M. Ito, Y. Kakuta, Biochem.
Biophys. Res. Commun. 374 (2008) 549–552.
[11] J.R. Ketudat Cairns, A. Esen, Cell. Mol. Life Sci. 67 (2010) 3389–3405.
[12] B. Cobucci-Ponzano, V. Aurilia, G. Riccio, B. Henrissat, P.M. Coutinho, A.
Strazzulli, A. Padula, M.M. Corsaro, G. Pieretti, G. Pocsfalvi, I. Fiume, R. Cannio,
M. Rossi, M. Moracci, J. Biol. Chem. 285 (27) (2010) 20691–20703.
[13] W. Schliemann, J. Plant Physiol. 116 (1984) 123–132.
[14] K.H. Lee, H.L. Piao, H.Y. Kim, S.M. Choi, F. Jiang, W. Hartung, I. Hwang, J.M.
Kwak, I.J. Lee, I. Hwang, Cell 126 (2006) 1109–1120.
[15] K. Kai, K. Wakasa, H. Miyagawa, Phytochemistry 68 (2007) 2512–2522.
[16] S. Wakuta, S. Hamada, H. Ito, H. Matsuura, K. Nabeta, H. Matsui,
Phytochemistry 71 (2010) 1280–1288.
[17] Z.Y. Xu, K.H. Lee, T. Dong, J.C. Jeong, J.B. Jin, Y. Kanno, D.H. Kim, S.Y. Kim, M. Seo,
R.A. Bressan, D.J. Yun, I. Hwang, Plant Cell 24 (2012) 2184–2199.
[18] Y. Koda, E.A. Omer, T. Yoshihara, H. Shibata, S. Sakamura, Y. Okazawa, Plant
Cell Physiol. 29 (1988) 1047–1051.
[51] B. Rempel, S. Withers, Glycobiology 18 (2008) 570–586.