Substrate specificity in hydrolysis and transglucosylation by family 1 β-glucosidases from cassava and Thai rosewood

Article abstract of DOI:10.1016/j.molcatb.2010.09.003

Thai rosewood (Dalbergia cochinchinensis Pierre) dalcochinase and cassava (Manihot esculenta Crantz) linamarase are glycoside hydrolase family 1 β-glucosidases with 47% amino acid sequence identity. Each enzyme can hydrolyze its natural substrate, dalcochinin-8′-O-β-d-glucoside and linamarin, respectively, but not the natural substrate of the other enzyme. Linamarase can transfer glucose to primary, secondary and tertiary alcohols with high efficiency, while dalcochinase can transglucosylate primary and secondary alcohols at moderate levels. In this study, eight amino acid residues in the aglycone binding pocket of dalcochinase were individually replaced with the corresponding residues of linamarase, in order to identify residues that may account for their catalytic differences. The residues I185 and V255 of dalcochinase appeared important for its substrate specificity, with their respective mutations resulting in 24- and 12-fold reductions in k _cat/K_m for the hydrolysis of dalcochinin-8′-O- β-d-glucoside. Transglucosylation activity was improved when I185, N189 and V255 of dalcochinase were replaced with A201, F205 and F271 of linamarase, respectively, suggesting these residues support transglucosylation in linamarase. Among these three mutants, only the N189F mutant showed significant increases in the rate constants for the reactivation of trapped glucosyl-enzyme intermediates by all alcohols. Together, our results suggest that both hydrophobicity and geometry are important determinants for substrate specificity in hydrolysis and transglucosylation by these family 1 β-glucosidases.

Full text of DOI:10.1016/j.molcatb.2010.09.003

Journal of Molecular Catalysis B: Enzymatic 67 (2010) 257–265
Contents lists available at ScienceDirect
Journal of Molecular Catalysis B: Enzymatic
journal homepage: www.elsevier.com/locate/molcatb
Substrate specificity in hydrolysis and transglucosylation by family 1
ˇ-glucosidases from cassava and Thai rosewood
Prachumporn T. Kongsaereea,b,∗, Khakhanang Ratananikoma, Khuanjarat Choengpanyab,
Nusra Tongtubtima,b, Penporn Sujiwattanaratb, Chompoonuth Porncharoennopb,
Amornrat Onpiuma, Jisnuson Svastic
a Department of Biochemistry, Faculty of Science, Kasetsart University, Bangkok 10900, Thailand
b Interdisciplinary Graduate Program in Genetic Engineering, Faculty of Graduate School, Kasetsart University, Bangkok 10900, Thailand
c Department of Biochemistry and Center for Excellence in Protein Structure and Function, Faculty of Science, Mahidol University, Bangkok 10400, Thailand
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 13 May 2010
Received in revised form 29 July 2010
Accepted 9 September 2010
Available online 17 September 2010
Thai rosewood (Dalbergia cochinchinensis Pierre) dalcochinase and cassava (Manihot esculenta Crantz)
linamarase are glycoside hydrolase family 1 ˇ-glucosidases with 47% amino acid sequence identity. Each
enzyme can hydrolyze its natural substrate, dalcochinin-8ꢀ-O-ˇ-d-glucoside and linamarin, respectively,
but not the natural substrate of the other enzyme. Linamarase can transfer glucose to primary, sec-
ondary and tertiary alcohols with high efficiency, while dalcochinase can transglucosylate primary and
secondary alcohols at moderate levels. In this study, eight amino acid residues in the aglycone bind-
ing pocket of dalcochinase were individually replaced with the corresponding residues of linamarase, in
order to identify residues that may account for their catalytic differences. The residues I185 and V255 of
dalcochinase appeared important for its substrate specificity, with their respective mutations resulting in
24- and 12-fold reductions in kcat/Km for the hydrolysis of dalcochinin-8ꢀ-O-ˇ-d-glucoside. Transglucosy-
lation activity was improved when I185, N189 and V255 of dalcochinase were replaced with A201, F205
and F271 of linamarase, respectively, suggesting these residues support transglucosylation in linamarase.
Among these three mutants, only the N189F mutant showed significant increases in the rate constants for
the reactivation of trapped glucosyl-enzyme intermediates by all alcohols. Together, our results suggest
that both hydrophobicity and geometry are important determinants for substrate specificity in hydrolysis
and transglucosylation by these family 1 ˇ-glucosidases.
Keywords:
Substrate specificity
ˇ-Glucosidase
Dalcochinase
Linamarase
Transglucosylation
© 2010 Elsevier B.V. All rights reserved.
1. Introduction
Dalcochinase is an isoflavonoid ˇ-glucosidase, purified from
the seeds of Thai rosewood (Dalbergia cochinchinensis Pierre) [3].
ˇ-Glucosidases (EC 3.2.1.21) catalyze the hydrolysis of the ˇ-O-
glycosidic linkage between a glucose moiety and an aglycone (aryl
or alkyl) or another sugar [1]. Most enzymes belong to glycoside
hydrolase families 1 and 3 (GH1 and GH3), and hydrolyze their
substrates via a double-displacement mechanism [2]. Isozymes
of ˇ-glucosidases differ considerably in their specificities for the
aglycone moiety linked to the glucosyl group, and show different
capabilities in catalyzing reverse hydrolysis and glucosyl transfer
reactions. These differences presumably result from specific inter-
actions between unique amino acid residues in their active site
pockets and their specific substrates.
Linamarase is a cyanogenic ˇ-glucosidase present in cassava (Mani-
hot esculenta Crantz) [4]. Both enzymes belong to GH1 family,
and contain the highly conserved TL/FNEP and I/VTENG motifs
that are present in all GH1 ˇ-glucosidases [5,6]. The two glu-
tamate residues, E182 and E396 in dalcochinase and E198 and
E413 in linamarase, present in the conserved motifs act as the
catalytic acid/base and nucleophile residues, respectively. Both
enzymes are predicted to have tertiary structures consisting of
an (˛/ˇ)8 barrel similar to other GH1 enzymes. While the two
enzymes share 47% sequence identity, their catalytic capabilities
are distinct. In hydrolysis, dalcochinase and linamarase can each
efficiently hydrolyze its own natural substrate, dalcochinin-8ꢀ-O-
ˇ-d-glucoside (Dal-Glc) and linamarin, respectively (Fig. 1), but
cannot hydrolyze the natural substrate of the other enzyme [7].
Furthermore, these two enzymes showed distinct capabilities in
transglucosylation and reverse hydrolysis reactions. Dalcochinase
can transglucosylate primary alcohols well, and secondary alco-
hols moderately, but not tertiary alcohols [8]. However, linamarase
Abbreviations: Dal-Glc, dalcochinin-8ꢀ-O-ˇ-d-glucoside; GH, glycoside hydro-
lase family; pNP-Glc, p-nitrophenyl-ˇ-d-glucoside.
∗
Corresponding author. Tel.: +66 2 562 5555x2026; fax: +66 2 561 4627.
E-mail address: fscippt@ku.ac.th (P.T. Kongsaeree).
1381-1177/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcatb.2010.09.003

258
P.T. Kongsaeree et al. / Journal of Molecular Catalysis B: Enzymatic 67 (2010) 257–265
2.2. Sequence alignment, homology modeling and molecular
docking
Since we have not been able to crystallize natural or recom-
binant dalcochinase for structure determination, despite efforts
over several years, residues likely to be involved in substrate bind-
ing were identified by molecular modeling. Alignment of amino
acid sequences of D. cochinchinensis (Thai rosewood) dalcochinase
(gi:6118076), M. esculenta (cassava) linamarase (gi:249261), Zea
mays (maize) ˇ-glucosidase ZmGlu1 (gi:13399866) and Sorghum
bicolor (sorghum) ˇ-glucosidase Dhr1 (gi:49259431) was per-
formed with ClustalW 2.0.10 [13]. The three-dimensional model
of wild-type dalcochinase was created with Geno3D [14], using
the structure of maize ˇ-glucosidase 1 complexed with a natu-
ral substrate DIMBOA-ˇ-d-glucoside (PDB code 1E56A [15]) as a
template (45% identity to dalcochinase). The overall structure of
the model was checked by the PROCHECK, ProSA, Verify-3D and
WHATIF programs [16–19]. The structure of Dal-Glc was built using
the Sybyl 7.2 molecular modeling package, and optimized with
the Tripos force field with 100 iterations of Simplex followed by
Powell minimization algorithm with a 0.05 kcal/mol energy gradi-
ent convergence criterion; partial atomic charges were attributed
with the Gasteiger–Huckel method (Tripos Inc., St. Louis, MO, USA).
The active site pocket of modeled enzymes was determined from
the coordinates of DIMBOA-ˇ-d-glucoside [15]. Dal-Glc was docked
into the active site pocket of modeled enzymes with GOLD 3.1 [20].
Fig. 1. Structures of dalcochinin-8ꢀ-O-ˇ-d-glucoside (a) and linamarin (b).
is the only enzyme studied thus far that can efficiently catalyze
transglucosylation reactions using primary, secondary and tertiary
alcohols as acceptors [9]. On the other hand, dalcochinase can also
synthesize oligosaccharides and glycosides by reverse hydrolysis
[10], while linamarase cannot. The distinct preferences shown by
Thai rosewood dalcochinase and cassava linamarase for the leav-
ing aglycone moieties in hydrolysis reactions and for the incoming
alcohol groups in transglucosylation reactions presumably result
from differences in the amino acids found in the aglycone binding
sites.
We have previously reported the cloning and expression of dal-
cochinase in the yeast Pichia pastoris [11]. Purified recombinant
dalcochinase exhibits enzymatic properties that are similar to those
of natural dalcochinase. This allows us to examine the identity
of amino acids conferring substrate specificity in hydrolysis and
transglucosylation in dalcochinase and linamarase via site-directed
mutagenesis. In this study, eight residues in the aglycone binding
pocket of dalcochinase were individually replaced with the corre-
sponding residues of linamarase. While detailed crystal structures
of both enzymes are still lacking, this report provides information
on the structure-function relationships of dalcochinase, and has
implications for other GH1 ˇ-glucosidases.
2.3. Construction of dalcochinase mutants
Eight single mutations were made via site-directed mutage-
nesis to replace amino acid residues in the aglycone binding
pocket of dalcochinase to the corresponding residues of linama-
rase. These mutations are I185A, N189F, M195V, H253F, V255F,
N323Q, A454N and E455I (the numbers indicate their positions in
the sequence of dalcochinase as reported in Ref. [5]). The recom-
binant plasmid pPICZ-His8-trncTRBG [11] was used as a template.
Site-directed mutagenesis was performed according to the method
published previously [21]. Each 50 ␮L mutagenesis reaction used a
sense/antisense mutagenic primer pair (only the sequences of the
sense mutagenic primers are shown in Table 1) and three units Pfu
DNA polymerase (Promega, Madison, WI, USA). The reactions were
then incubated with 10 units DpnI (Promega) at 37 ◦C overnight
to remove the parental double-stranded DNA. The DpnI-treated
DNA was transformed into competent E. coli by electroporation,
and selected on LB-agar plates containing 25 ␮g/mL zeocin at 37 ◦C.
The plasmids containing correctly mutated dalcochinase sequences
were identified by DNA sequencing. Subsequently, the mutant
plasmids were linearized with SacI, transformed into P. pastoris
by electroporation, and selected on YPDS plates with 100 ␮g/mL
zeocin, following the protocols from Invitrogen.
2. Experimental
2.1. Materials
Escherichia coli strain DH5␣ was used as a cloning host. P.
pastoris strain GS115 [his4] (Invitrogen, Carlsbad, CA, USA) was
used as an expression host. The recombinant plasmid pPICZ-His8-
trncTRBG, containing the coding sequence of dalcochinase with
an N-terminal truncation following the ␣ mating factor propep-
tide and 8 histidine residues [11], was used for the expression
of recombinant wild-type dalcochinase. Natural dalcochinase, nat-
ural linamarase, Dal-Glc and linamarin were purified from their
natural sources as described previously [3,7,12]. p-Nitrophenyl-
ˇ-d-glucoside (pNP-Glc), 2ꢀ,4ꢀ-dinitrophynyl 2-deoxy-2-fluoro-ˇ-
d-glucoside, 4-methylumbelliferyl-ˇ-d-glucoside, and glucose-
oxidase/peroxidase mixture were purchased from Sigma Chemical
(St. Louis, MO, USA). 2,2ꢀ-Azonobis-3-ethylbenz-thiazolinesulfonic
acid was purchased from Roche (Mannheim, Germany). Silica gel 60
F254 aluminum sheets for TLC were purchased from Merck (Darm-
stadt, Germany).
2.4. Expression and purification of recombinant dalcochinase in
P. pastoris
Colonies of P. pastoris harboring wild-type and mutant con-
structs of dalcochinase were grown in BMGH with 4 × 10−3% (w/v)
histidine and 4 × 10−5% (w/v) biotin overnight at 30 ◦C, 200 rpm.
The culture was changed into BMMH with 4 × 10−3% (w/v) his-
tidine, 4 × 10−5% (w/v) biotin and 0.5% (w/v) casamino acid, and
induced with 0.5% (v/v) methanol every day. Recombinant dal-
cochinase (wild-type and mutant) was secreted into the culture
media, and its ˇ-glucosidase activity was assayed every 2 days until
stable. Then, recombinant dalcochinase was purified from the cul-
ture media by hydrophobic interaction chromatography followed
by immobilized metal-ion affinity chromatography as described
previously [11].

P.T. Kongsaeree et al. / Journal of Molecular Catalysis B: Enzymatic 67 (2010) 257–265
259
Table 1
Sequences of the sense mutagenic primers (5ꢀ → 3ꢀ) used in the generation of eight single mutants. The underlined bases indicated the mutation site. Sequences of the
antisense mutagenic primers are the reverse complements of the sequences shown.
Dalcochinase mutant
Primer sequences
I185A
GGATTACACTAAATGAGCCATCAGCTTTCACCGCGAATGGGTATGC
GAGCCATCAATCTTCACCGCGTTTGGGTATGCATACGG
GGGTATGCATACGGTGTTTTTGCACCAGGTCGATGTTCTCC
CAGAAAGGTACAATAGGCATTTCCTTGTTTGTAGTTTGGGTTATACCGC
GGCATTTCCTTGCACGTATTTTGGGTTATACCGC
GGGTTCATTTGATTTTATTGGACTACAATATTACACCACTAACTATGCTACC
GGTCATTGTTGGACAACTTTGAATGGAATGAGGGTTATACATCACG
CATTGTTGGACAACTTTGAATGGGCTATTGGTTATACATCACGATTTGG
N189F
M195V
H253F
V255F
N323Q
A454N
E455I
Protein concentration was determined by the Bio-Rad Pro-
tein Assay (Bio-Rad, Hercules, CA, USA). SDS-PAGE (7.5% resolving
gel) was performed to check the purity of protein [22]. West-
ern blot analysis and chemiluminescent detection was done with
mouse monoclonal antibody against natural dalcochinase (a gift
from Dr. Watchara Kasinrerk, Chiangmai University, Thailand),
horseradish peroxidase-conjugated rabbit polyclonal antibody
against mouse immunoglobulins (Dako, Glostrup, Denmark) and
ECL Plus Western Blotting Detection reagents (GE Healthcare,
Buckinghamshire, UK). Non-denaturing PAGE and activity staining
with 1 mM 4-methylumbelliferyl-ˇ-d-glucoside was performed as
reported previously [3].
with shaking. The alcohols tested were methanol, ethanol, 1-
propanol, 2-propanol, 2-methyl-1-propanol, 2-methyl-2-propanol,
1-butanol and 2-butanol. The reaction was stopped by boiling for
2 min. Aliquots of the reactions (8 ␮L each) with the same alco-
hol acceptors were spotted on the same TLC sheet. The TLC was
developed twice in ethyl acetate/methanol/water (16:6:1 by vol)
for 4.25 cm and twice in 2-propanol/ethanol/water (5:1:2 by vol)
for 1.75 cm, and visualized by dipping into 20% (v/v) sulfuric acid
in ethanol and incubating at 125 ◦C for 10 min. All TLC plates con-
tained standard markers, consisting of 20, 40, 60 and 80 nmol of
glucose, methyl glucoside and pNP-Glc. The chromatograms were
scanned with a GS-8000 imaging densitometer (Bio-Rad) and the
amounts of alkyl glucoside products were quantified with Quality
One version 4.2.1 software. Standard curves between density and
amounts of standard markers with correlation coefficient (r2) of at
least 0.95 were used. Results were expressed in percent of moles of
alkyl glucoside out of the total moles of free glucose, pNP-Glc and
alkyl glucoside present in each reaction.
The rate constants of reactivation of the trapped glucosyl-
enzyme intermediates by various alcohols were determined
following the methods described previously [23]. Purified enzymes
were inactivated with 50 ␮M 2ꢀ,4ꢀ-dinitrophynyl 2-deoxy-2-
fluoro-ˇ-d-glucoside at 30 ◦C for 30 min in 0.1 M sodium acetate,
pH 5.0, giving a trapped 2-deoxy-2-flouro-glucosyl-enzyme inter-
mediate. The excess inhibitor was removed by centrifugal
ultrafiltration, and fresh buffer was added to make up the same
protein concentration. The trapped inactive enzyme intermediate
was then reactivated by incubating in 0.1 M sodium acetate, pH
5.0, alone (hydrolysis), or in the presence of 0.5 M alcohols (trans-
glucosylation) at 30 ◦C. The regained activity was measured by the
hydrolysis of 200 ␮M 4-methylumbelliferyl-ˇ-d-glucoside, and the
release of 4-methylumbelliferone was monitored by a continuous
fluorometric spectrophotometer (excitation wavelength = 350 nm
and emission wavelength = 450 nm). The observed rate constants
of reactivation (kre,obs) of each enzyme by a particular alcohol was
determined from the slope of the plot of ln[(V0 − V)/V0] versus time
of incubation with alcohol, when V and V0 represent the initial
velocity of the reaction in the presence and absence of inhibitor,
respectively.
2.5. Hydrolytic activities and kinetic measurements
ˇ-Glucosidase activity of recombinant dalcochinase (wild-type
and mutant) in the culture media and during purification was
assayed with 1 mM pNP-Glc in 0.1 M sodium acetate, pH 5.0, in
a 0.5 mL reaction at 30 ◦C for 30 min. The reaction was stopped by
adding 1 mL 2 M sodium carbonate, pH 10, and the p-nitrophenol
released was measured by its absorbance at 400 nm.
To determine the level of enzyme activity of purified enzymes
for kinetic measurement and transglucosylation reactions, ˇ-
glucosidase activity was assayed with 15 mM pNP-Glc in 0.1 M
sodium acetate, pH 5.0, in a 0.1 mL reaction at 30 ◦C for 5 min. The
reaction was stopped by adding 1 mL 2 M sodium carbonate, pH 10,
and p-nitrophenol released was measured at 400 nm. One unit of
activity is defined as the amount of enzyme used to release 1 ␮mol
of p-nitrophenol in 1 min.
Kinetic parameters for the hydrolysis of pNP-Glc, Dal-Glc and
linamarin were determined by incubating 0.1 unit of purified
enzymes with substrates at various concentrations (0–45 mM
pNP-Glc and linamarin, or 0–25 mM Dal-Glc) in 0.1 M sodium
acetate, pH 5.0, in a 50 ␮L reaction at 30 ◦C for 5 min. The reac-
tion was stopped by boiling for 5 min and immediately cooled
to room temperature. Subsequently, the reaction was incubated
with 50 ␮L 2 mg/ml 2,2ꢀ-Azonobis-3-ethylbenz-thiazolinesulfonic
acid and 100 ␮L 5 unit/ml glucose oxidase reagent for 15 min at
37 ◦C, and glucose released was calculated by comparison of the
absorbance at 410 nm to that from a glucose standard curve. Kinetic
parameters (Km, kcat and kcat/Km) were calculated by nonlinear
regression of the Michaelis–Menten equation with KaleidaGraph
(Synergy Software, Reading, PA, USA).
3. Results and discussion
3.1. Sequence alignment, homology modeling and molecular
docking
2.6. Transglucosylation activities and rate constants of
reactivation by alcohols
Sequence alignment and structural modeling were performed in
order to identify amino acid residues in the aglycone binding pock-
ets that differ between dalcochinase and linamarase. The amino
acid sequences of dalcochinase and linamarase were aligned with
those of maize ˇ-glucosidase ZmGlu1 and sorghum ˇ-glucosidase
Dhr1 (Fig. 2). A three-dimensional homology model of dalcochinase
was generated with the crystal structure of maize ˇ-glucosidase
ZmGlu1 in complex with its natural substrate DIMBOA-ˇ-d-
Transglucosylation activities of various enzymes were assessed
by quantification of alkyl glucoside formation on TLC as described
previously [8]. Purified enzymes (0.1 unit) were incubated with
10 mM pNP-Glc (acting as a glucosyl donor) and 0.9 M alkyl alco-
hols of chain length C1–C4 (acting as a glucosyl acceptor) in 0.1 M
sodium acetate, pH 5.0, in a 100 ␮L reaction for 20 h at 30 ◦C

260
P.T. Kongsaeree et al. / Journal of Molecular Catalysis B: Enzymatic 67 (2010) 257–265
TRDC
-----IDFAKEVRETI------TEVPP-FNRSCFPSDFIFGTASSSYQYEGE----GRVP 44
MLVLFISLLALTRPAMGTDDDDDNIPDDFSRKYFPDDFIFGTATSAYQIEGEATAKGRAP 60
--SARVGSQ-NGVQMLSP----SEIPQ---RDWFPSDFTFGAATSAYQIEGAWNEDGKGE 50
--AQTISSESAGIHRLSP----WEIPR---RDWFPPSFLFGAATSAYQIEGAWNEDGKGP 51
CVLM
ZmGlu1
SbDhr1
TRDC
SIWDNFTHQYPEKIADRSNGDVAVDQFHRYKKDIAIMKDMNLDAYRMSISWPRILPTGRV 104
SVWDIFSKETPDRILDGSNGDVAVDFYNRYIQDIKNVKKMGFNAFRMSISWSRVIPSGRR 120
SNWDHFCHNHPERILDGSNSDIGANSYHMYKTDVRLLKEMGMDAYRFSISWPRILPKGTK 110
STWDHFCHNFPEWIVDRSNGDVAADSYHMYAEDVRLLKEMGMDAYRFSISWPRILPKGTL 111
CVLM
ZmGlu1
SbDhr1
TRDC
SGGINQTGVDYYNRLINESLANGITPFVTIFHWDLPQALEDEYGGFLNHS---VVNDFQD 161
REGVNEEGIQFYNDVINEIISNGLEPFVTIFHWDTPQALQDKYGGFLSRD---IVYDYLQ 177
EGGINPDGIKYYRNLINLLLENGIEPYVTIFHWDVPQALEEKYGGFLDKSHKSIVEDYTY 170
AGGINEKGVEYYNKLIDLLLENGIEPYITIFHWDTPQALVEAYGGFLDER---IIKDYTD 168
CVLM
ZmGlu1
SbDhr1
▼
▼
▼
TRDC
YADLCFQLFGDRVKHWITLNEPSIFTANGYAYGMFAPGRCSPSYNPTCTGGDAGTETYLV 221
YADLLFERFGDRVKPWMTFNEPSAYVGFAHDDGVFAPGRCSSWVNRQCLAGDSATEPYIV 237
FAKVCFDNFGDKVKNWLTFNEPQTFTSFSYGTGVFAPGRCSPGLDCAYPTGNSLVEPYTA 230
FAKVCFEKFGKTVKNWLTFNEPETFCSVSYGTGVLAPGRCSPGVSCAVPTGNSLSEPYIV 228
CVLM
ZmGlu1
SbDhr1
□
●
●
▼ ▼
TRDC
AHNLILSHAATVQVYKRKYQEHQKGTIGISLHVVWVIPLSNSTSDQNATQRYLDFTCGWF 281
CVLM
AHNLLLSHAAAVHQYRKYYQGTQKGKIGITLFTFWYEPLSDSKVDVQAAKTALDFMFGLW 297
ZmGlu1
SbDhr1
GHNILLAHAEAVDLYNKHYKRDD-TRIGLAFDVMGRVPYGTSFLDKQAEERSWDINLGWF 289
AHNLLRAHAETVDIYNKYHKGAD-GRIGLALNVFGRVPYTNTFLDQQAQERSMDKCLGWF 287
□
○ ○
▼
TRDC
MDPLTAGRYPDSMQYLVGDRLPKFTTDQAKLVKGSFDFIGLNYYTTNYATKSDASTCCPP 341
MDPMTYGRYPRTMVDLAGDKLIGFTDEESQLLRGSYDFVGLQYYTAYYAEPIPPVDPKFR 357
LEPVVRGDYPFSMRSLARERLPFFKDEQKEKLAGSYNMLGLNYYTSRFSKNIDISPNYSP 349
LEPVVRGDYPFSMRVSARDRVPYFKEKEQEKLVGSYDMIGINYYTSTFSKHIDLSPNNSP 347
□
CVLM
ZmGlu1
SbDhr1
TRDC
SYLTDPQVTLLQ--QRNGVFIGPVTPSGWMCIYPKGLRDLLLYFKEKYNNPLVYITENGI 399
RYKTDSGVNATPY-DLNGNLIGPQAYSSWFYIFPKGIRHFLNYTKDTYNDPVIYVTENGV 416
VLNTDDAYASQEVNGPDGKPIGPPMGNPWIYMYPEGLKDLLMIMKNKYGNPPIYITENGI 409
VLNTDDAYASQETKGPDGNAIGPPTGNAWINMYPKGLHDILMTMKNKYGNPPMYITENGM 407
●
CVLM
ZmGlu1
SbDhr1
▼▼
TRDC
DEKNDAS--LSLEESLIDTYRIDSYYRHLFYVRYAIR-SGANVKGFFAWSLLDNFEWAEG 456
CVLM
DNYNNES--QPIEEALQDDFRISYYKKHMWNALGSLKNYGVKLKGYFAWSYLDNFEWNIG 474
ZmGlu1
SbDhr1
GDVDTKETPLPMEAALNDYKRLDYIQRHIATLKESID-LGSNVQGYFAWSLLDNFEWFAG 468
GDIDKGD--LPKPVALEDHTRLDYIQRHLSVLKQSID-LGADVRGYFAWSLLDNFEWSSG 464
■■
●●
○
TRDC
YTSRFGLYFVNYTT-LNRYPKLSATWFKYFLARDQESAKLEILAPKARWSLSTMIKEEKT 515
YTSRFGLYYVDYKNNLTRYPKKSAHWFTKFLNISVNANNIYELTSKDSRKVGKFYVM--- 531
FTERYGIVYVDRNNNCTRYMKESAKWLKEFN-TAKKP-SKKILTPA-------------- 512
YTERFGIVYVDRENGCERTMKRSARWLQEFNGAAKKVENNKILTPAGQLN---------- 514
CVLM
ZmGlu1
SbDhr1
TRDC
KPKRGIEGF 524
---------
CVLM
ZmGlu1
SbDhr1
---------
---------
Fig. 2. Alignment of amino acid sequences of Thai rosewood dalcochinase, cassava linamarase and maize ˇ-glucosidase. The alignment is generated with ClustalW 2.0.10
using the reported sequences of Thai rosewood dalcochinase (TRDC, gi:6118076), cassava linamarase (CVLM, gi:249261) maize ˇ-glucosidase Glu1 (ZmGlu1, gi:13399866)
and sorghum ˇ-glucosidase Dhr1 (SbDhr1, gi:49259431). The sequence of Thai rosewood dalcochinase is shown without the first 23 N-terminal residues, MLAMTSKAILLL-
GLLALVSTSAS, predicted to be a signal peptide [5]. The highly conserved (L/F)NEP and (I/V)TENG motifs are underlined, and the two active site glutamate residues are shown
in BOLD. The residues that were targeted for site-directed mutagenesis are marked with ꢀ above the sequences. The residues shown to be located in the active site or involved
in aglycone binding in dalcochinase, cassava linamarase, maize ˇ-glucosidase Glu1 and sorghum ˇ-glucosidase Dhr1 are marked with ꢁ, ꢂ, ᭹ and ꢁ below the sequences,
respectively [15,24–26].
glucoside (PDB code 1E56A) as a template. This structure showed
the highest identity (45%) to dalcochinase among all structures of
ˇ-glucosidases with complexed natural substrates, which would
simplify subsequent docking studies.
The quality of the homology model was evaluated by the
PROCHECK, ProSA, Verify-3D and WHATIF programs. PROCHECK
showed 97% of the residues in the homology model were located
in the most favorable or allowed regions of the Ramachandran
plot and only 1.5% of them in the disallowed region (compared
with 98.5% and 0%, respectively, for the template structure). The
G-factor values in PROCHECK (in which a score above −0.5 indi-
cates a reliable model) were obtained to be 0.22 and 0.34 for the
model and the template structures, respectively. The ProSA z-scores
of the selected dalcochinase model and the template structure were
−8.15 and −9.63, respectively, which were within the acceptable
range. Assessment of the compatibility of the residues with their
environment by Verify-3D (in which a score over 0.2 were con-
sidered reliable) showed 91.2% of residues in the model had a
score over 0.2, compared with 93.9% for the template. The pack-
ing quality of each residue was also evaluated by the WHATIF
program (in which a score above −5.0 is deemed satisfactory).
In this case, the selected dalcochinase model showed an average
score of −1.357 (95.7% of residues with scores above −5), while the
template structure obtained an average score of −0.774 (98.0% of
residues with scores above −5). The root mean square deviation
˚
of the C␣ atoms was calculated to be 1.68 A with respect to the

P.T. Kongsaeree et al. / Journal of Molecular Catalysis B: Enzymatic 67 (2010) 257–265
261
linamarase, namely A201, F205, V211, F269, F271, Q339, N472 and
I473, respectively. Amino acid residues corresponding to all of these
positions in other GH1 enzymes, except M195 in dalcochinase,
have been previously shown to be involved in enzyme–substrate
interactions. Crystal structures of maize ˇ-glucosidase ZmGlu1 [15]
and sorghum ˇ-glucosidase Dhr1 [25] in complex with their nat-
ural substrates, DIMBOA-ˇ-d-glucoside and dhurrin, respectively,
have shown the amino acid residues corresponding to N189, H253,
V255, A454 and E455 of dalcochinase included in this study, to
be in direct contact with the aglycone groups of the substrates.
Previous mutational analyses in linamarase [24] and dalcochinase
[26] have identified the amino acid residues corresponding to I185,
H253, N323, A454 and E455 of dalcochinase included in this study,
as essential active site residues for catalysis and substrate speci-
ficity. The residue corresponding to M195 of dalcochinase, which
falls within the motif GM/VFAPGRCS between residues 194 and
202 of dalcochinase, is replaced with V in linamarase, maize ˇ-
glucosidase ZmGlu1 and sorghum ˇ-glucosidase Dhr1. As the side
chain of M195 pointed toward the aglycone group of Dal-Glc in our
model, its mutation to V was included in this study.
Fig. 3. Three-dimensional model of the active site pocket of dalcochinase with a
docked molecule of Dal-Glc. The homology model of wild-type dalcochinase was
created with Geno3D [14], using the structure of maize ˇ-glucosidase 1 complexed
with a natural substrate DIMBOA-ˇ-d-glucoside (PDB code 1E56A [15]) as a tem-
plate. The modelisshownasviewedfromtheexterioroftheenzyme, looking intothe
active site pocket, with the glucose moiety buried behind the structure of the agly-
cone moiety. The main chain conformation is shown in gray ribbon. The side chains
that are targeted for site-directed mutagenesis are shown as green stick models.
Dal-Glc is shown as yellow ball-and-stick model. The catalytic residues (E182 and
E396) and the conserved residues (F196 and W368) are shown in gray. The picture
was generated with PyMOL version 0.99 (DeLano Scientific, Palo Alto, CA, USA).
3.2. Construction and production of dalcochinase mutants
Eight single mutants of dalcochinase, which are I185A, N189F,
M195V, H253F, V255F, N323Q, A454N and E455I, were constructed
via site-directed mutagenesis. All wild-type and mutant forms of
dalcochinase were expressed in P. pastoris, with levels of total ˇ-
glucosidase activity in culture media at the time of harvest varying
between 9.2 units (I185A) and 102 units (M195V) per 1-L culture.
Purification yields were between 15 and 40%, and specific activities
ranged from 0.15 unit/mg (I185A) to 1.30 unit/mg (H253F).
All wild-type and mutant forms of dalcochinase showed a broad
band with an apparent molecular mass of about 66 kDa on SDS-
PAGE (Supplementary data 1, left panels), which migrated more
slowly than the apparent 63 kDa band of natural dalcochinase. The
broadening effect and the greater apparent molecular weight of
recombinant dalcochinase are most likely due to greater glycosyla-
tion in yeast than in plants [27]. Indeed, we have previously shown
that the natural and recombinant wild-type dalcochinase contained
approximately 13.6 and 16.1% sugar (by weight), respectively [11].
All wild-type and mutant forms of dalcochinase could be detected
equally well by a mouse monoclonal antibody against natural dal-
cochinase (Supplementary data 1, right panels), and migrated to the
same distance on non-denaturing gels (Supplementary data 2, left
panels), suggesting that the overall conformation of the wild-type
enzyme was retained in the mutants. However, they exhibited vari-
able activities toward 1 mM 4-methylumbelliferyl-ˇ-d-glucoside
template. All evaluations suggest that a reliable homology model
for Thai rosewood dalcochinase has been obtained for subsequent
docking study. The model showed an (␣/␤)8 barrel structure as
expected for GH1 enzymes (not shown). The two catalytically active
site residues (E182 and E396 in dalcochinase) were separated by
˚
about 5 A (as measured between the O␧1 atoms of both residues)
and were located near the bottom of the binding pocket, which
˚
˚
appeared as a 25 A long and 7 A wide slot-like pocket.
The structure of Dal-Glc, the natural substrate of dalcochinase,
was then docked into the active site pocket of the homology model
of dalcochinase (Fig. 3). A fitness score of docked Dal-Glc into the
dalcochinase model was obtained to be 61.41. Re-docking DIMBOA-
ˇ-d-glucoside into the active site pocket of maize ˇ-glucosidase
ZmGlu1 was performed as a control docking, and gave a fitness
score of 69.15. In the algycone binding pocket of the homology
model, the position of the Dal-Glc was directed mainly through
stacking ꢀ–ꢀ interactions between ring B (phenyl ring) of Dal-Glc
and the indole ring of the conserved W368. There appeared to be a
hydrophobic interaction of Dal-Glc with I185. Also, the methylene
group at C-7ꢀ of Dal-Glc appeared to fit between H253 and W368
of dalcochinase. Further inside the active site pocket, N323 inter-
actedwith thecarboxylgroupsofbothacid/basecatalyst (E182) and
nucleophile (E396). Q339, the corresponding position in linama-
rase, has been shown via site-directed mutagenesis to coordinate
the catalytic diad [24]. At the mouth of the pocket, Rings D and E of
Dal-Glc appeared almost perpendicular to rings A, B and C, and were
exposed to solvent, with no apparent specific interaction with any
residues of dalcochinase. The distances between Dal-Glc and amino
acid residues in binding pocket of dalcochinase model as predicted
by molecular docking were summarized in Table 2.
Table 2
Distances between Dal-Glc and amino acid residues in binding pocket of dalcochi-
nase model as predicted by molecular docking.
Dal-Glc position
Amino acid position
Distance (Å)
Sugar ring
O2
O3
O4
O6
Y325-HH
2.49
2.89
1.48
2.16
1.72
N181-HD22
W453-H␧1
W445-H␧1
E396-O␧1
H at O2
Cꢀ
Methylene
Methylene
Ring-A
H
Ring-B
Ring-B
Ring-C
OH
H235-H␧1
W368-Hz3
2.34
2.51
From the sequence alignment and structural modeling, eight
amino acid residues in the aglycone binding pocket of dalcochi-
nase, namely I185, N189, M195, H253, V255, N323, A454 and E455,
were chosen as targets for site-directed mutagenesis to investi-
gate their roles in directing substrate specificity. Each residue in
dalcochinase would be replaced with the corresponding residue in
I185-HD11
W368
2.41
␲–␲ (3.52–4.09)
W368-H␧1
2.35

262
P.T. Kongsaeree et al. / Journal of Molecular Catalysis B: Enzymatic 67 (2010) 257–265
Table 3
Kinetic parameters for the hydrolysis of pNP-Glc and Dal-Glc of wild-type and mutant forms of dalcochinase and linamarase. kcat values were estimated assuming a subunit
molecular weight of 66,000.
Enzymes
pNP-Glc
Dal-Glc
Km (mM)
kcat (s−1
)
kcat/Km (s−1mM−1
)
Km (mM)
kcat (s−1
)
kcat/Km (s−1mM−1
)
Wild-type dalcochinase
Natural linamarase
I185A
N189F
M195V
H253F
V255F
N323Q
A454N
3.91 0.31
0.72 0.04
14.33 1.07
0.59 0.07
1.77 0.16
2.69 0.30
2.67 0.17
1.84 0.18
1.34 0.13
11.83 0.88
237.0 5.2
131.9 1.3
12.3 0.4
172.5 2.9
36.4 0.6
58.0 1.7
10.6 0.1
36.2 0.6
157.9 3.1
49.8 1.3
60.6 5.0
183.2 10.3
0.86 0.07
292.4 35.0
20.6 1.9
21.6 2.5
4.0 0.3
19.7 2.0
117.8 11.7
4.2 0.3
5.67 0.33
N.D.a
62.3 1.7
N.D.a
11.0 0.7
N.D.a
1.62 0.21
0.63 0.05
6.25 0.76
4.35 0.42
1.74 0.16
4.94 0.49
3.63 0.60
4.69 0.45
0.74 0.04
36.0 0.9
31.4 2.2
21.9 0.9
1.5 0.1
33.5 1.5
14.7 1.4
120.7 5.2
0.46 0.06
57.1 4.7
5.0 0.7
5.0 0.5
0.89 0.09
6.8 0.7
4.0 0.8
25.7 2.7
E455I
a
N.D., no activity detected.
in activity staining of non-denaturing gels (Supplementary data 2,
right panels), reflecting altered activities as a result of mutations.
in kcat/Km, respectively. So, the reduction in substrate specificity
toward Dal-Glc by these two mutations, as judged by the values of
kcat/Km, appeared to result from reduced catalytic activity rather
than changes in binding affinity. These two mutations also showed
large decreases in the kcat for the hydrolysis of pNP-Glc. A201 has
been shown to be in the active site of linamarase [24], and here
we showed the role of the corresponding position in dalcochi-
nase, I185, in directing substrate specificity. Molecular modeling
of Dal-Glc in the active site pocket of dalcochinase suggested a
hydrophobic contact between side chain of I185 and the agly-
cone moiety of Dal-Glc (Fig. 3). The I185A mutation could create
a void space in the substrate binding pocket of dalcochinase lead-
ing to incorrect positioning of the aglycone moiety and placing the
glucosidic bond in an unfavorable orientation for hydrolysis. Sim-
ilar reasons were proposed for the reduction in kcat of F193A and
F193I mutations in Zm-p60.1 maize ˇ-glucosidase, corresponding
to position N189 of dalcochinase in this study, for the hydroly-
sis of pNP-Glc [28,29]. In addition, the I185A mutation resulted in
a decreased hydrophobic environment close to the catalytic site
that might affect the pKas of the catalytic glutamates. Indeed, an
A201V mutation (increasing hydrophobicity) in linamarase, corre-
sponding to position I185 of dalcochinase in this study, increased
the value of pKe2 from 7.22 to 7.44 for the hydrolysis of pNP-Glc
[24]. The possibilities of mis-positioning of the glucosidic bond for
hydrolysis and changing the pKas of the catalytic glutamates may
also account for the reduction in kcat caused by the V255F mutation.
Other interesting mutants, N189F and E455I, showed unex-
pected 5- and 2-fold increases in substrate specificity for Dal-Glc,
respectively, as judged by the values of kcat/Km, compared with
the recombinant wild-type dalcochinase. The increase in kcat/Km of
the N189F mutant was due to a 9-fold reduction in Km, suggest-
ing favorable hydrophobic interaction between phenylalanine and
Dal-Glc at this position. On the other hand, the increase in kcat/Km
of the E455I mutant was due to a 2-fold increase in kcat. So, it
appeared that increased hydrophobicity as a result of isoleucine
substitution at this position may lead to better placement of the
substrate for hydrolysis. Four other mutations led to 2- to 3-fold
reductions in kcat/Km, as a result of 2- to 4-fold reductions in kcat
but small changes in Km, compared with the recombinant wild-
type enzyme, indicating that their substitutions affected catalysis
rather than binding. However, it should be noted that a reduction
in Km may not always correspond to improved substrate binding,
but may also result from a reduced efficiency of the deglycosylation
step by water in a double-displacement mechanism.
3.3. Kinetic measurement for hydrolytic activities
The kinetic properties of recombinant wild-type dalcochinase,
dalcochinase mutants and natural linamarase for the hydrolysis
of pNP-Glc and Dal-Glc were determined (Table 3). The recom-
binant wild-type dalcochinase showed kinetic properties toward
pNP-Glc and Dal-Glc that are similar to those reported for natural
dalcochinase [3,7]. Cassava linamarase hydrolyzed pNP-Glc with
a 6-fold lower Km and a 2-fold lower kcat, resulting in a 3-fold
higher efficiency compared with the recombinant wild-type dal-
cochinase, suggesting that residues in the substrate binding pocket
of linamarase are more suited for the binding of pNP-Glc than those
in dalcochinase. Only linamarase could hydrolyze linamarin with
a Km of 0.84 0.07 mM, a kcat of 79.8 1.6 s−1, and a kcat/Km of
95 8 s−1 mM−1. In agreement with previous studies, neither the
recombinant wild-type dalcochinase could appreciably hydrolyze
linamarin, nor could linamarase hydrolyze Dal-Glc [7].
In hydrolysis of pNP-Glc, the I185A and E455I mutations
showed 3-fold increases in Km, and 70- and 14-fold decreases in
kcat/Km, respectively, compared with the recombinant wild-type
enzyme. Presumably, I185 and E455 in dalcochinase could form
hydrophobic and ionic interactions to the pNP moiety of pNP-
Glc, respectively, so they were more favorable for the binding
and hydrolysis of pNP-Glc than the corresponding positions in
linamarase, A201 and I473. On the other hand, the N189F and
A454N mutations yielded 5- and 2-fold increases in kcat/Km, respec-
tively, resulting from 7- to 3-fold decreases in Km, respectively,
and a 1.5-fold decrease in kcat, compared with the recombinant
wild-type enzyme. So their corresponding positions in linamarase,
F205 and N472, may contribute to better binding to pNP-Glc com-
pared with dalcochinase by offering suitable hydrophobic and polar
interactions with the substrate, respectively. Four other mutations
in dalcochinase resulted in lower hydrolytic efficiencies (from 3-
to 15-fold) compared with the recombinant wild-type enzyme,
mostly due to 4- to 22-fold reductions in the kcat. Among all
mutants, V255F showed the greatest decrease in kcat (22-fold) com-
pared with the recombinant wild-type enzyme, but little change in
Km, suggesting that F271 in linamarase could disturb the hydrolysis
but not binding of pNP-Glc.
In hydrolysis of Dal-Glc, the natural substrate of dalcochinase,
all dalcochinase mutants exhibited less than an order of mag-
nitude reduction in Km compared to the recombinant wild-type
enzyme, indicating small effects of mutations on the binding prop-
erties. On the other hand, two mutants, I185A and V255F, showed
84- and 40-fold decreases in kcat compared with the recombi-
nant wild-type dalcochinase, resulting in 24- and 12-fold decreases
However, none of single mutants of dalcochinase could
hydrolyze linamarin. Combinations of some of these mutations in
the aglycone binding pocket may offer cumulative effects that are
needed to overcome steric hindrance imposed by the three sub-
stituents of the 1-cyano-1-methylethyl moiety of linamarin. Also,

P.T. Kongsaeree et al. / Journal of Molecular Catalysis B: Enzymatic 67 (2010) 257–265
263
Table 4
Mole percent of alkyl glucoside products from transglucosylation reactions of wild-type and mutant forms of dalcochinase and linamarase. Each value is an average from
duplicate TLC plates. Alkyl glucosides were quantitated using methyl glucoside as standard.
Enzymes
Primary alcohols
Secondary alcohols
Tertiary alcohol
Methanol
Ethanol
1-Propanol
1-Butanol
2-Methyl-1-propanol
2-Propanol
2-Butanol
2-Methyl-2-propanol
Wild-type dalcochinase
Natural linamarase
I185A
N189F
M195V
H253F
V255F
N323Q
A454N
22
52
20
43
24
29
40
29
29
26
42
51
40
64
46
44
54
43
50
40
68
73
73
86
68
66
80
69
68
67
79
71
92
96
90
87
92
88
85
87
71
74
90
90
71
75
84
74
75
80
4
60
18
7
3
6
15
6
4
12
65
25
11
10
11
21
12
15
10
N.D.a
35
N.D.a
N.D.a
N.D.a
N.D.a
N.D.a
N.D.a
N.D.a
N.D.a
E455I
7
a
N.D., no product detected.
other amino acid residues further away from the aglycone binding
pocket may contribute to favorable environment for the hydrolysis
of linamarin.
using primary and secondary alcohols suggested that transglu-
cosylation may be improved by creating a more hydrophobic
environment (in the case of the N189F mutation), and providing
suitable geometry for enzyme–substrate interactions (in the cases
of the I185A and V255F mutations). Thus, the positions A201, F205
and F271 in linamarase were favorable for transglucosylation activ-
ity. This is in agreement with their significant roles in substrate
binding in cassava linamarase, maize ˇ-glucosidase ZmGlu1 and
sorghum ˇ-glucosidase Dhr1, respectively [15,24,25].
To evaluate the transglucosylation activities more accurately,
the method of Hommalai et al. [23] was applied to test the
I185A, N189F and V255F mutants that showed higher transglu-
cosylation activities than the recombinant wild-type enzyme. The
trapped 2-deoxy-2-flouro-glucosyl-enzyme intermediates were
reactivated by the transfer of a 2-deoxy-2-flouro-glucosyl moiety
from the enzymes to the alcohol acceptors. In all cases, the plots
of ln[(V0 − V)/V0] versus time of incubation with alcohol were lin-
ear, suggesting pseudo-first-order reactions (an example is shown
in Supplementary data 4). The observed rate constants of enzyme
reactivation (kre,obs) in the absence and presence of alcohol rep-
resented the progress of the hydrolysis and transglucosylation
reactions, respectively (Table 5).
Under the conditions used, our values of kre,obs for the recom-
binant wild-type dalcochinase were similar to those obtained
previously for the natural enzyme with all alcohols tested, except
for 1-butanol and 2-methyl-1-propanol, where our values were
about half of those reported [23]. Of the dalcochinase mutants, the
I185A mutant gave statistically significant increases in the values
of kre,obs compared with the recombinant wild-type enzyme with
only primary alcohol acceptors and 2-propanol, while the V255F
mutant showed significant increases in the values of kre,obs with
only 1-butanol and 2-methyl-1-propanol.
However, the N189F mutant showed statistically significant
increases in the values of kre,obs with all alcohols tested com-
pared with the recombinant wild-type enzyme. Surprisingly, this
mutant also gave a 3-fold increase in the value of kre,obs with 2-
methyl-2-propanol, compared to its own hydrolysis reaction and
to the recombinant wild-type dalcochinase, indicating improved
transglucosylation activity to tertiary alcohol acceptor. This lower
effectiveness of water compared to alcohol in cleaving the glucosyl-
enzyme, might also be partly responsible for the lower Km value
observed in hydrolysis of pNP-Glc and Dal-Glc with the N189F
mutant. However, the kre,obs with 2-methyl-2-propanol was still
very much lower than that reported for cassava linamarase
(47 6 × 10−4 min−1) [23], which was 8-fold faster than simple
hydrolysis. The transglucosylation rate constants and dissociation
constants of alcohols were not determined in this study since most
kre,obs values were less than 10-fold different from the rates of
hydrolysis.
3.4. Transglucosylation activities and rate constants of
reactivation by alcohols
In the transglucosylation reactions, the enzyme cleaves pNP-Glc
to form a glucosyl-enzyme intermediate complex and transfers the
glucose to an alcohol acceptor, yielding alkyl glucoside, detectable
as a spot on TLC. Free glucose could result from transfer of glu-
cose to water, or secondary hydrolysis of alkyl glucoside products.
The spots of alkyl glucoside products and glucose varied in size
and intensity. An example of TLC from transglucosylation reac-
tions with methanol as acceptors is shown in Supplementary data
3. It should be noted that transglucosylation is a kinetic process,
and the amounts of alkyl glucoside product may vary with time.
However, previous studies showed that in reactions of dalcochi-
nase, hydrolysis of pNP-Glc took place rapidly with little hydrolysis
of alkyl glucosides over 24 h [8]. So, in this study, only single-
point measurements of transglycosylation yield were performed in
order to provide quick assessment of the transglucosylation activ-
ities of dalcochinase mutants toward various alcohols, using an
incubation time of 20 h to maximize cleavage of pNP-Glc donor
and minimize secondary hydrolysis of alkyl glucoside products.
The mole percents of alkyl glucoside products from all trans-
glucosylation reactions are shown in Table 4. The recombinant
wild-type dalcochinase showed transglucosylation activities to pri-
mary and secondary alcohol acceptors that are comparable to
those previously reported for the natural enzyme [8]. In agree-
ment with previous studies, linamarase was the only enzyme that
could transfer glucose to the tested primary, secondary as well
as tertiary alcohol acceptors [9,23]. However, our product yields
from linamarase were lower than those previously reported [9],
probably due to lower starting concentrations of pNP-Glc and alco-
hols.
For transglucosylation reactions to primary alcohols, only the
N189F and V255F mutants showed improved product yields com-
pared with the recombinant wild-type dalcochinase with all
primary alcohols tested, while the I185A mutant gave better yields
only with longer chain primary alcohols. For transglucosylation to
secondary alcohols, only the I185A and V255F mutants were bet-
ter than the wild-type enzyme. However, none of our mutants
could use a tertiary alcohol as a glucosyl acceptor, indicating
that single mutations were insufficient to change the architec-
ture of the aglycone binding pocket of dalcochinase to accept
aglycone groups with three substituents at the carbinol carbon,
such as the 1-cyano-1-methylethyl moiety of linamarin and 2-
methyl-2-propanol. Nonetheless, our transglucosylation results

264
P.T. Kongsaeree et al. / Journal of Molecular Catalysis B: Enzymatic 67 (2010) 257–265
Table 5
Observed rate constants of enzyme reactivation (kre,obs × 10−4 min−1) of wild-type and mutant forms of dalcochinase for a series of 0.5 M alcohols. The reported values are
averages from three independent experiments with standard deviations.
Alcohols
Wild-type dalcochinase
I185A
3.7 0.2a
N189F
4.7 0.5a
V255F
Primary alcohols
Methanol
Ethanol
1-Propanol
1-Butanol
2-Methyl-1-propanol
Secondary alcohols
2-Propanol
2.4 0.3
3.1 0.2
4.6 0.8
6.3 1.1
5.4 1.0
1.7 0.3
1.6 0.5
3.9 0.8
7.5 0.3a
7.7 1.4a
3.6 0.1a
7.3 0.2a
9.7 0.8a
20.4 2.4a
6.9 0.9a
36.0 7.8a
98.9 6.5a
84.9 4.7a
2.5 0.5
2.4 0.3
3.6 0.2a
2.7 0.1
8.1 1.0a
1.5 0.4
2.5 0.3
2-Butanol
12.2 2.1a
Tertiary alcohol
2-Methyl-2-propanol
Control (no acceptor)
2.7 0.3
2.2 0.1
3.1 0.3
2.6 0.3
8.9 1.3a
3.4 0.7a
2.3 0.3
1.3 0.4
a
Statistically significant increase compared to the recombinant wild-type dalcochinase (p-value ≤ 0.05).
Discrepancies between the TLC results and the values of kre,obs
may partly be due to secondary product hydrolysis that occurred
during the 20-h incubation in the former method. In some reac-
tions, the mutant enzymes gave lower alkyl glucoside yields, but
exhibited significantly higher kre,obs values compared with the
recombinant wild-type dalcohinase, suggesting better alkyl gluco-
side cleavage by the mutant than the wild-type enzyme. On the
other hand, in some reactions, the mutant enzymes gave higher
alkyl glucoside yields, but exhibited similar or lower kre,obs values
compared with the recombinant wild-type dalcohinase, suggesting
lower alkyl glucoside cleavage by the mutant than the wild-type
enzyme.
Since the proton removal in the deglycosylation step appeared
to be the major step governing transition state formation in
the transglucosylation reaction of linamarase [23], we also used
1,1,1-trifluoromethyl-2-propanol, which has a similar structure to
2-methyl-2-propanol but a lower pKa value (11.6 versus 18.0), to
reactivate the trapped 2-deoxy-2-flouro-glucosyl-enzyme inter-
mediates of all enzymes. However, the resulting kre,obs values of
the wild-type dalcochinase and all three mutants did not differ sig-
nificantly from simple hydrolysis (pKa of water ∼15.7) (results not
shown). So, the nucleophilic strength appeared less important than
the complementarity of size, shape and the hydrophobicity of the
incoming alcohol acceptor and the aglycone binding pocket of the
enzyme.
complementarity in both geometry and hydrophobicity of the sub-
strate and the binding pocket of enzyme.
Acknowledgements
The authors especially thank Palangpon Kongsaeree for help
in the preparations of Dal-Glc and linamarin, Patchreenart
Saparpakorn for help with modeling programs, and Watchara
Kasinrerk for providing the mouse monoclonal antibody against
natural dalcochinase. The comments and suggestions from James
Ketudat-Cairns during the preparation of manuscript are much
appreciated. The project was financially supported by the Thailand
Research Fund and the Commission on Higher Education (Grant
numbers MRG4980131 to P.T.K. and MRG485S015 to P.S.), and by
the Faculty of Science, Kasetsart University. K.R. and K.C. are recip-
ients of the Ph.D. scholarships from the Commission on Higher
Education, Thailand.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.molcatb.2010.09.003.
References
[1] A. Esen, in: A. Esen (Ed.), ␤-Glucosidases Biochemistry and Molecular Biology,
American Chemical Society, Washington, DC, 1993, pp. 1–14.
[2] B. Henrissat, G.D. Davies, Plant Physiol. 124 (2000) 1515–1519.
[3] C. Srisomsap, J. Svasti, R. Surarit, V. Champattanachai, K. Boonpuan, P. Sawan-
gareetrakul, P. Subhasitanont, D. Chokchaichamnankit, J. Biochem. 119 (1996)
585–590.
[4] Y. Eksittikul, M. Chulavatanatol, Arch. Biochem. Biophys. 266 (1988) 263–269.
[5] J.R. Ketudat Cairns, V. Champattanachai, C. Srisomsap, B. Wittman-Liebold, B.
Thiede, J. Svasti, J. Biochem. 128 (2000) 999–1008.
[6] M.A. Hughes, K. Brown, A. Pancoro, B.S. Murray, E. Oxtoby, J. Hughes, Arch.
Biochem. Biophys. 295 (1992) 273–279.
[7] J. Svasti, C. Srisomsap, S. Techasakul, R. Surarit, Phytochem 50 (1999) 739–743.
[8] K. Lirdprapamongkol, J. Svasti, Biotechnol. Lett. 22 (2000) 1889–1894.
[9] J. Svasti, T. Phongsak, R. Sarnthima, Biochem. Biophys. Res. Commun. 305 (2003)
470–475.
[10] C. Srisomsap, P. Subhasitanont, S. Techasakul, R. Surarit, J. Svasti, Biotechnol.
Lett. 21 (1999) 947–951.
[11] P. Toonkool, P. Metheenukul, S. Sujiwattanarat, P. Paiboon, N. Tongtubtim,
M. Ketudat-Cairns, J. Ketudat-Cairns, J. Svasti, Protein Expr. Purif. 48 (2006)
195–204.
[12] P.M. Dewick, in: J.B. Harborne (Ed.), The Flavonoids: Advances in Research Since
1986, Chapman and Hall, London, 1993, pp. 117–238.
[13] D.G. Higgins, J.D. Thompson, T.J. Gibson, Nucleic Acids Res. 22 (1994)
4673–4680.
4. Conclusions
Taken together, our mutational analysis in the aglycone bind-
ing pocket of dalcochinase has identified the amino acid residues
that may help explain the distinct aglycone specificities in hydroly-
sis and transglucosylation. In particular, residues I185 and V255 of
dalcochinase play important roles in substrate specificity as judged
by decreases in the kcat/Km ratios of their mutants for hydrolyzing
Dal-Glc. On the other hand, residues N189 and E455 of dalcochinase
may not be optimal for Dal-Glc hydrolysis, as their mutations led
to increases in the kcat/Km ratios. Since both side chains were polar,
our results suggested that these positions might prefer hydropho-
bic residues. Alkyl glucoside synthesis by transglucosylation was
improved by the I185A, N189F and V255F mutations, suggesting
the positive roles played by the corresponding residues, A201,
F205 and F271 of linamarase. However, only the N189F mutant
showed significant increases in the reactivation rate constants with
all alcohols, compared to the wild-type enzyme, while the other
two mutants showed improvement only with longer-chain primary
alcohols. So it appeared that the yield of alkyl glucoside products
depended on the efficiencies of both glucose transfer and secondary
product cleavage. Our results suggest that substrate specificity in
hydrolysis and transglucosylation is accomplished as a result of
[14] C. Combet, M. Jambon, G. Deléage, C. Geourjon, Bioinformatics 18 (2002)
213–214.
[15] M. Czjzek, M. Cicek, V. Zamboni, D.R. Bevan, B. Henrissat, A. Esen, Proc. Natl.
Acad. Sci. U.S.A. 97 (2000) 13555–13560.
[16] R.A. Laskowski, M.W. MacArthur, D.S. Moss, J.M. Thronton, J. Appl. Crystallogr.
26 (1993) 283–291.
[17] M. Wiederstein, M.J. Sipp, Nucleic Acids Res. 35 (2007) 407–410.

P.T. Kongsaeree et al. / Journal of Molecular Catalysis B: Enzymatic 67 (2010) 257–265
265
[18] R. Luthy, J.U. Bowie, D. Eisenberg, Nature 356 (1992) 83–85.
[19] G. Vriend, J. Mol. Graph. 8 (1990) 52–56.
[20] M.L. Verdonk, J.C. Cole, M.J. Hartshorn, C.W. Murray, R.D. Taylor, Proteins 52
(2003) 609–623.
[21] J. Braman, C. Papworth, A. Greener, Methods Mol. Biol. 57 (1996) 31–44.
[22] U.K. Laemmli, Nature 227 (1970) 680–685.
[23] G. Hommalai, P. Chaiyen, J. Svasti, Arch. Biochem. Biophys. 442 (2005) 11–20.
[24] Z. Keresztessy, K. Brown, M.A. Dunn, M.A. Hughes, Biochem. J. 353 (2001)
199–205.
[26] P. Chuankhayan, T. Rimlumduan, W. Tantanuch, N. Mothong, P.T. Kongsaeree, P.
Metheenukul, J. Svasti, O.N. Jensen, J.R. Ketudat Cairns, Arch. Biochem. Biophys.
468 (2007) 205–216.
[27] J.L. Cereghino, J.M. Cregg, FEMS Microbiol. Rev. 24 (2000) 45–66.
[28] R. Dopitová, P. Mazura, L. Janda, R. Chaloupková, P. Jerˇábek, J.
Damborsky´, T. Filipi, N.S. Kiran, B. Brzobohaty´, FEBS J. 275 (2008) 6123–
6135.
[29] J. Zouhar, J. Vévodová, J. Marek, J. Damborsky´, X.D. Su, B. Brzobohaty´, Plant
Physiol. 127 (2001) 973–985.
[25] L. Verdoucq, J. Morinière, D.R. Bevan, A. Esen, A. Vasella, B. Henrissat, M. Czjzek,
J. Biol. Chem. 279 (2004) 31796–31803.