Rice BGlu1 glycosynthase and wild type transglycosylation activities distinguished by cyclophellitol inhibition

Article abstract of DOI:10.1016/j.carres.2012.02.013

The rice BGlu1 β-d-glucosidase nucleophile mutant E386G is a glycosynthase that catalyzes the synthesis of cellooligosaccharides from α-d-glucopyranosyl fluoride (GlcF) donor and p-nitrophenyl (pNP) cellobioside (Glc2-pNP) or cello-oligosaccharide accepto

Full text of DOI:10.1016/j.carres.2012.02.013

Carbohydrate Research 352 (2012) 51–59
Contents lists available at SciVerse ScienceDirect
Carbohydrate Research
journal homepage: www.elsevier.com/locate/carres
Rice BGlu1 glycosynthase and wild type transglycosylation activities
distinguished by cyclophellitol inhibition
Salila Pengthaisong a, Chi-Fan Chen b, Stephen G. Withers b, Buabarn Kuaprasert c,
James R. Ketudat Cairns a,
⇑
a Schools of Biochemistry and Chemistry, Institute of Science, Suranaree University of Technology, 111 University Avenue, Muang District, Nakhon Ratchasima 30000, Thailand
b Centre for High-throughput Biology (CHiBi) and Department of Chemistry, University of British Columbia, 2036 Main Mall, Vancouver, British Columbia, Canada V6T 1Z1
c Research Facility Division, Synchrotron Light Research Institute (Public Organization), 111 University Avenue, Muang District, 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:
The rice BGlu1 b-
of cellooligosaccharides from
side (Glc2-pNP) or cello-oligosaccharide acceptors. When activity with other donors and acceptors was
tested, the initial enzyme preparation cleaved pNP-b- -glucopyranoside (Glc-pNP) and pNP-b- -fucopyr-
D
-glucosidase nucleophile mutant E386G is a glycosynthase that catalyzes the synthesis
Received 1 December 2011
Received in revised form 6 February 2012
Accepted 10 February 2012
Available online 21 February 2012
a
-D-glucopyranosyl fluoride (GlcF) donor and p-nitrophenyl (pNP) cellobio-
D
D
anoside (Fuc-pNP) to pNP and glucose and fucose, suggesting contamination with wild type BGlu1 b-glu-
cosidase. The products from reaction of GlcF and Fuc-pNP included Fuc-b-(1?3)-Fuc-pNP, Glc-b-(1?3)-
Fuc-pNP, and Fuc-b-(1?4)-Glc-b-(1?3)-Fuc-pNP, suggesting the presence of both wild type BGlu1 and
its glycosynthase. Inhibition of the BGlu1 b-glucosidase activity within this preparation by cyclophellitol
confirmed that the E386G glycosynthase preparation was contaminated with wild type BGlu1. Rice BGlu1
E386G-2, generated from a new construct designed to minimize back-mutation, showed glycosynthase
activity without wild type hydrolytic or transglycosylation activity. E386G-2 catalyzed transfer of glyco-
Keywords:
Glycosylation
Glycosynthase
Oligosaccharide synthesis
Rice
Transglucosylation
b-Glucosidase
syl residues from GlcF,
syl fluoride, and -xylosyl fluoride donors to Glc2-pNP acceptor. The synthetic products from the
reactions of -fucosyl fluoride and -mannosyl fluoride donors were confirmed to result from addition
of a b-(1?4)-linked glycosyl residue. Moreover, the E386G glycosynthase transferred glucose from GlcF
donor to glucose, cellobiose, Glc-pNP, Fuc-pNP, pNP-b- -galactopyranoside, and pNP-b- -xylopyranoside
acceptors, but little to pNP-b- -mannopyranoside. Production of longer oligosaccharides occurred most
a-L-arabinosyl fluoride, a-D-fucosyl fluoride, a-D-galactosyl fluoride, a-D-manno-
a
-D
a
a
D
D
D
readily on acceptors with an equatorial 4-OH. Elimination of wild type contamination thereby allowed
a clear assessment of BGlu1 E386G glycosynthase catalytic abilities.
Ó 2012 Elsevier Ltd. All rights reserved.
1. Introduction
using 4-nitrophenyl b-
transfer to pyridoxine or Glc-pNP acceptors, resulting in the syn-
thesis of pyridoxine 50-O-b-
-glucoside and pNP gluco-oligosaccha-
rides. Rice BGlu1 has previously been mutated to form
D-glucoside (Glc-pNP) as donor for glucosyl
Glycosynthases are nucleophile mutants of glycosidases that
effect synthesis of oligosaccharides and glycoconjugates from
glycosyl fluoride donors and suitable acceptors without hydrolysis
of the products (Fig. 1).1–3 Glycosynthases derived from endoglyco-
sidases use oligosaccharyl fluorides as donor substrates and have
high regioselectivity.3–8 Glycosynthases derived from exoglycosid-
ases are typically capable of efficient glycoside assembly with
monosaccharyl fluoride donors, but sometimes with lower
substrate specificity and regioselectivity.3,9,10
D
a
glycosynthase (Fig. 1C) that can synthesize pNP-cellooligosaccha-
rides of at least 11 b-(1?4)-linked glucosyl residues.14 In that
study, it was found that the E386G mutation (designated E414G
in the previous study, based on the nucleophile position in the
BGlu1 precursor) resulted in much higher glycosynthase activity
than the E386A or E386S mutations. The ability to synthesize long
cellooligosaccharides was attributed to their binding in a long
oligosaccharide binding cleft, which has recently been verified in
the X-ray crystal structures of BGlu1 acid-base and glycosynthase
mutants.15,16
In this paper we expand our exploration of the glycosynthase
activity of the BGlu1 E386G enzyme, in particular, we explore
the ability of the enzyme to utilize alternate glycosyl fluoride
donors and acceptors, thereby broadening the spectrum of
di- and trisaccharides produced by this glycosynthase. Wild type
A
glycoside hydrolase family 1 b-D-glucosidase from rice
(BGlu1, systematically called Os3BGlu7) has a broad substrate
specificity, with particularly high hydrolytic activity toward
b-(1?3)- and b-(1?4)-linked oligosaccharides and pyridoxine
50-O-b- -glucoside.11–13 It also has high transglucosylation activity
D
⇑
Corresponding author. Tel.: +66 44 224304; fax: +66 44 224185.
E-mail address: cairns@sut.ac.th (J.R. Ketudat Cairns).
0008-6215/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
doi:10.1016/j.carres.2012.02.013

52
S. Pengthaisong et al. / Carbohydrate Research 352 (2012) 51–59
(Glc-pNP), pNP b-
ranoside (Gal-pNP), pNP-b-
b- -xylopyranoside (Xyl-pNP), and pNP
D
-fucopyranoside (Fuc-pNP), pNP-b-
-mannopyranoside (Man-pNP), pNP
-arabinopyranoside, as
D-galactopy-
D
D
a-L
well as b-(1?2), b-(1?3), b-(1?4), and b-(1?6) glycosidic link-
ages.11,12 This suggests that BGlu1 glycosynthase might also utilize
other glycosyl fluoride donors and thus make mixed products.
When the acceptor specificity of the BGlu1 E386G glycosynth-
ase preparation was investigated, both transglycosylation and
hydrolysis products were detected. The Glc-pNP and Fuc-pNP
acceptors were cleaved to pNP, glucose and fucose, while Glc-
pNP was produced from Glc2-pNP, which suggested wild type con-
tamination of the BGlu1 E386G glycosynthase preparation. Since
new column packings were used for all glycosynthase prepara-
tions, cross-contamination during purification was unlikely. It thus
appeared that a mutation in the Origami(DE3) Escherichia coli
strain being used for expression may have led to a higher than nor-
mal reversion rate, possibly by misreading on the ribosome or
reversion mutations in a few cells in the expression culture.18,19
This suggested that some of the products formed may have been
generated by a combination of glycosynthase and wild type trans-
glycosylation activities. Characterization of purified products pro-
vided further evidence for the combined action of BGlu1 wild
type and E386G glycosynthase, as described below.
Mass spectrometric and NMR characterizations identified the
purified products of reaction of
a-glucosyl fluoride (GlcF) and
Fuc-pNP catalyzed by the BGlu1 E386G preparation as Fuc-b-
(1?3)-Fuc-pNP, Glc-b-(1?3)-Fuc-pNP, and Fuc-b-(1?4)-Glc-b-
(1?3)-Fuc-pNP (Table 1, products 1, 2, and 3, respectively). Since
glycosynthase mutants of BGlu1 show very little pNP-glycoside
hydrolysis,14 the nonreducing end fucosyl moieties in products of
Fuc-b-(1?3)-Fuc-pNP and Fuc-b-(1?4)-Glc-b-(1?3)-Fuc-pNP
from reaction of GlcF and Fuc-pNP may have come from transfu-
cosylation from Fuc-pNP by wild type BGlu1 contamination, as dia-
grammed in Figure 1A.
2.2. Cyclophellitol inhibition
The presence of wild type BGlu1 contamination in the E386G
preparation was confirmed by inhibition of the wild type enzyme
with the covalent inhibitor cyclophellitol. As shown in Figure 1B,
cyclophellitol inactivates retaining b-D-glucosidases by forming a
covalent adduct with the catalytic nucleophile. This is a stable ad-
duct that does not undergo reactivation by transglycosylation.20
Figure 2 shows that inactivation of wild type rice BGlu1 by cyclop-
hellitol follows pseudo-first order kinetics of inactivation, similar
to previous reports for other b-glucosidases.20,21 Cyclophellitol acts
as an irreversible inactivator of the rice BGlu1 b-glucosidase with a
Ki of 5.7 l
M and a ki of 0.35 minꢀ1 (Fig. 2B and C).
The E386G preparation was thus pre-incubated with cyclophell-
itol for 2 h, which resulted in complete inhibition of wild type
BGlu1 activity (Fig. 2A), before testing the transglycosylation of
GlcF onto Fuc-pNP. The E386G glycosynthase preparation synthe-
sized two products in the range of pNP-disaccharides and a few tri-
saccharides in the reaction without cyclophellitol pre-incubation
(Fig. 3 lane EG ꢀ) and a single product, corresponding to Glc-
Fuc-pNP, after pre-incubation with 2 mM cyclophellitol (Fig. 3 lane
EG +).
Figure 1. Reaction mechanisms of hydrolysis and transglycosylation catalyzed by
rice BGlu1 b-glucosidase (A); covalent intermediate trapping of rice BGlu1 by
cyclophellitol (B), and rice BGlu1 E386G glycosynthase (C). Note that cyclophellitol
cannot react with BGlu1 E386G due to its lack of a catalytic nucleophile.
BGlu1 contamination of BGlu1 E386G was confirmed and elimi-
nated by inhibition with cyclophellitol, thereby allowing the actual
glycosynthase products to be identified.
As expected, untreated wild type BGlu1 b-glucosidase synthe-
sized transglycosylation products, while that pre-incubated with
cyclophellitol did not (Fig. 3, lanes BG). Moreover, uninhibited wild
type BGlu1 synthesized only one product in the pNP-disaccharide
range: Fuc-Fuc-pNP, which was also seen with the untreated
E386G preparation, but not when it was preincubated with cyclop-
hellitol. This result confirms that the original E386G glycosynthase
was contaminated with wild type BGlu1, and that, as would be ex-
pected, the wild type BGlu1 does not transfer from GlcF, suggesting
2. Results and discussion
2.1. Evidence of wild type BGlu1 contamination in the BGlu1
E386G preparation
As is the case with many GH1 glycosidases,17 the wild type
BGlu1 b-
D
-glucosidase has relatively broad substrate specificity,
-glucopyranoside
since it can hydrolyze 4-nitrophenyl b-
D

S. Pengthaisong et al. / Carbohydrate Research 352 (2012) 51–59
53
Table 1
Structures of products of glycosynthase reactions of BGlu1 E386G
#
Donor
Acceptor
Product
1
2
3
⁄
⁄⁄Formed by wild type BGlu1 b-glucosidase contaminant.
Formed by successive action of glycosynthase and wild type BGlu1 contaminant.
it is unlikely to have contributed to the production of hetero-disac-
charide products where the added glycosyl residues were derived
from an a-glycosyl fluoride donor.
were preferred on Glc residues with equatorial 4-OH groups, but
the glycosyl residue was added onto the equatorial 3-OH of Fuc
residues, where the 4-OH is axial.
The cyclophellitol inhibition results show that the combination
of wild type BGlu1 and E386G glycosynthase seen in the contami-
nated prep was likely responsible for synthesis of the novel pNP-
oligosaccharide Fuc-b-(1?4)-Glc-b-(1?3)-Fuc-pNP, which could
not be produced by either enzyme alone. This points to the impor-
tance of testing for wild type hydrolase activity when assessing
glycosynthase products, since the transglycosylation products aris-
ing from small amounts of contaminating hydrolase can be signif-
icant, given the large amount of glycosynthase needed for the
transfer reactions.
Streptomyces b-glucosidase glycosynthase is regiospecific for
production of b-(1?3)-linkages to pNP-monosaccharide acceptors
and can use Fuc-pNP and Gal-pNP acceptors,9 while Agrobacterium
b-glucosidase has a strong preference for formation of b-(1?4)-
linkages.22–25 As such, rice BGlu1 appears to be more similar to
Streptomyces b-glucosidase glycosynthase in its activity toward
these acceptors.
2.4. Transglycosylation activity of E386G-2
A new construct for the production of BGlu1 E386G was made
by changing the codon used from GAA (Glu) to GGC instead of
GGG to give Gly at the nucleophile position. Use of this double-
changed codon should decrease the probability of misreading as
Glu (GAG) by the ribosome or of a reversion mutation in the
expression culture. The protein expressed from this mutation
(E386G-2) indeed exhibited 33-fold lower Glc-pNP hydrolysis
activity compared to the original E386G glycosynthase (the
Glc-pNP hydrolysis activities of BGlu1, BGlu1 E386G, and BGlu1
E386G-2 were 1900, 1.1, and 0.033 nmol minꢀ1 mgꢀ1 protein,
respectively). When the E386G-2 glycosynthase preparation was
used to catalyze the reaction between Fuc-pNP and GlcF, only a
single product, corresponding to Glc-Fuc-pNP, was detected after
pre-incubation either without or with 2 mM cyclophellitol (Fig. 3,
lanes EG2 ꢀ and +).
2.3. Verification of linkages of glycosynthase products
As noted above, the wild type BGlu1 b-glucosidase could not
transfer glycosyl residues from
a-glycosyl fluoride donors, so any
products generated from such transfers reflect the glycosynthase
regiospecificity, even when derived from the mixture with wild
type in the E386G preparation. It was previously demonstrated
that reaction of BGlu1E386G with GlcF and Glc2-pNP specifically
yields b-(1?4)-linked oligosaccharides.14 The MS and NMR analy-
ses described above identified a product generated from transfer of
a single glucosyl residue onto Fuc-pNP as b-
pNP (Table 1, product 2). Similarly, in reaction products with
Glc2-pNP as acceptor, a single transfer product from -mannosyl
D-Glc-(1?3)-b-D-Fuc-
a
-D
fluoride (ManF) was identified as Man-b-(1?4)-Glc-b-(1?4)-
Glc-pNP (Table 1, product 4) and the corresponding product with
Since no wild type hydrolysis and transglycosylation activities
were detected in the E386G-2 preparation under the conditions
of the standard reaction, this sample was used to test the BGlu1
a-
D-fucosyl fluoride (FucF) donor was Fuc-b-(1?4)-Glc-b-(1?4)-
Glc-pNP (Table 1, product 5). Thus, in each case, b-(1?4)-linkages

54
S. Pengthaisong et al. / Carbohydrate Research 352 (2012) 51–59
acceptors were employed, similar to what had been seen with gly-
cosynthases from Agrobacterium b-glucosidase.22,23 By contrast, the
Streptomyces E383A glycosynthase could only synthesize disaccha-
rides when using GlcF and pNP-monosaccharide acceptors.9 When
using Fuc-pNP as acceptor E386G-2 synthesized only Glc-Fuc-pNP
disaccharide and a tiny amount of Glc-Glc-Fuc-pNP trisaccharide,
and showed no hydrolytic activity toward acceptors, verifying that
the longer Fuc-pNP oligosaccharides seen with the original E386G
prep resulted from the wild type BGlu1 transglycosylation activity.
The pNP-disaccharide product spot seen when Glc-pNP was used
as an acceptor was very faint (Fig. 4B, lane pG), explaining why
we failed to detect it in our previous work.14 This spot had a mobil-
ity slightly lower than Glc2-pNP, suggesting it might be pNP-gen-
tiobioside, but this identity was not confirmed due to the small
amount produced. Similarly, the product spots obtained with cello-
biose as acceptor were very faint, though at least two products
could be detected by sulfuric acid charring (Fig. 4D, lane C2).
Surprisingly, a clear spot of laminaribiose was observed when
glucose was used as an acceptor, and this spot was also seen in
the other reactions, likely due to the presence of glucose from
hydrolysis of GlcF. The production of laminaribiose is in line with
the fact that laminaribiose is a much better hydrolytic substrate
for the wild-type BGlu1 b-glucosidase than are cellobiose or other
disaccharides12 and consistent with structural studies in com-
plexes with laminaribiose.15 The enzyme also catalyzed autocon-
densation of GlcF, as seen by the smear above glucose in the TLC
and LC–MS peaks with masses of 379 m/z ([M+35Cl]ꢀ).
Transglycosylation yields for reaction of BGlu1 E386G-2 with
GlcF and a variety of acceptors are shown in Table 2, and those
with a variety of donors and Glc2-pNP acceptor are shown in Table
3. Relatively high yields were obtained with FucF and ManF donors,
but only a low yield with XylF. The absence of a substituent at C-5
apparently affects binding at the ꢀ1 subsite, although the activity
was higher than that of non-evolved glycosynthases derived from
Agrobacterium b-glucosidase22–24 and Streptomyces E383A,9 for
which XylF was not a donor substrate. Moreover, E386G-2 uses
AraF as a donor, something which has not previously been reported
for a GH1 glycosynthase. The E386G-2 glycosynthase gave low
yields with Fuc-pNP, Gal-pNP, Glc-pNP, and Xyl-pNP acceptors,
and GlcF donor, compared to the Streptomyces E383A.9
Although Gal-pNP and Fuc-pNP acceptors gave the highest
yields for disaccharide production, only very small amounts of tri-
saccharides and no tetrasaccharides were detected on LC–MS (Table 2).
In contrast, Glc-pNP gave higher production of trisaccharide and
tetrasaccharide products were also seen for Glc-pNP, Xyl-pNP,
and Man-pNP, which all have equatorial 4-OH groups. This agrees
with the previous report that the BGlu1 E386G glycosynthase ap-
pears to require a b-(1?4)-linked acceptor, such as Glc2-pNP or
cellotriose, before it can efficiently transfer successive glucosyl res-
idues onto a b-(1?4)-linked chain with high regioselectivity.14
With Glc2-pNP as acceptor, the BGlu1 E386G-2 glycosynthase
gave lower yields of trisaccharide products (26% with GlcF), but
higher yields of tetrasaccharide (16%) and longer products, com-
pared to Agrobacterium E358A (79% trisaccharide and 13% tetrasac-
charide under comparable conditions) and E358S (88%
trisaccharide, 7% tetrasaccharide) and Streptomyces E383A glyco-
synthase (80% trisaccharide and 3% tetrasaccharide). As noted pre-
viously, the unique feature of rice BGlu1 glycosynthases is their
ability to produce long oligosaccharides,14 so that BGlu1 E386G
also produced cellopentaose in 11% yield. At donor to acceptor ra-
tios above 1:1, the yields could not be assessed by HPLC with GlcF
donor and Glc2-pNP acceptor, since they resulted in precipitated
products. Thus, BGlu1 E386G is more useful for synthesis of longer
oligosaccharides, while the bacterial enzymes are better for syn-
thesis of di- and tri-saccharides.
Figure 2. Cyclophellitol inactivation of rice BGlu1. (A) Time course for incubation of
the enzymes with cyclophellitol. The enzyme solutions of 0.2
l
M BGlu1 (N), 40
lM
E386G (ꢀ), and 40 M E386G-2 (j) were pre-incubated with 1
l
l
M cyclophellitol
for different times and then assayed for release of pNP from Glc-pNP. (B) Semi-
logarithmic plot of wild type BGlu1 b-glucosidase activity versus time at various
concentrations of cyclophellitol, (d), 0.5
(ꢀ), 1.5 M. (C) Replot of the inverse of the apparent first order rate constants (kapp
from B versus the inverse of the inhibitor concentration, 1/[I].
lM; (h), 0.75 lM; (N), 1 lM; (j), 1.25 lM;
l
)
E386G glycosynthase activity with various donors and acceptors.
The E386G-2 glycosynthase successfully transferred a range of do-
nors (
a-
L-arabinosyl fluoride (AraF), FucF,
a-D-galactosyl fluoride
(GalF), GlcF, ManF, and
a-D-xylosyl fluoride (XylF)) onto pNP-cello-
biose to produce trisaccharides, and a faint spot of tetrasaccharide
was seen with FucF donor (Fig. 4).
The E386G-2 glycosynthase could also use glucose, cellobiose,
Fuc-pNP, Gal-pNP, Glc-pNP, Xyl-pNP, and Glc2-pNP as acceptors
(Fig. 4). pNP-Oligosaccharides resulting from addition of more than
one Glc could only be detected by TLC with Glc2-pNP as acceptor.
Longer oligosaccharides were, however, synthesized by E386G-2
using GlcF donor when Glc-pNP, Man-pNP, Xyl-pNP, or Glc2-pNP

S. Pengthaisong et al. / Carbohydrate Research 352 (2012) 51–59
55
Figure 3. Effect of cyclophellitol inhibition on transglycosylation products of rice BGlu1, BGlu1 E386G, and BGlu1 E386G-2 with Fuc-pNP and GlcF. The 0.2 lM BGlu1, 40 lM
E386G, and 40
lM E386G-2 were pre-incubated without (ꢀ) or with (+) 2 mM cyclophellitol for 2 h, then transglycosylation reactions with 10 mM Fuc-pNP and 10 mM a-
GlcF in 150 mM NH4HCO3, pH 7.0, were incubated at 30sC for 16 h. Silica gel TLC analysis was performed with EtOAc–MeOH–water (7:2.5:1) as a developing solvent, and
products detected under UV light (A) or by dabbing with 10% sulfuric acid in ethanol followed by charring (B). Lane G, glucose standard; GF, GlcF; C, control of 10 mM Fuc-pNP
and 10 mM GlcF without enzyme; BG, BGlu1; EG, BGlu1 E386G; EG2, BGlu1 E386G-2.
2.5. Conclusion
3.3. Transglycosylation activity of E386G with various donors
The E386G glycosynthase was incubated at 40 M with 10 mM
Pure rice BGlu1 E386G glycosynthase transfers glycosyl resi-
l
dues from a variety of
a
-glycosyl fluoride donors, in the reactivity
of each of the donors: AraF, FucF, GalF, ManF, or XylF and 10 mM
Glc2-pNP acceptor in 150 mM ammonium bicarbonate buffer, pH
7.0, at 30 °C for 16 h.14 Products were analyzed by TLC (Silica Gel
60 F254, aluminum-backed, Merck) with ethyl acetate (EtOAc)–
methanol (MeOH)–water (7:2.5:1) as solvent. Plates were visual-
ized under ultraviolet (UV) light and by exposure to 10% sulfuric
order Glc > Fuc > Ara > Man > Gal > Xyl. Likewise it transfers gluco-
syl residues from GlcF to several different acceptors, with a prefer-
ence for formation of 1–4 linkages. Only equatorial hydroxyls on
the ring appear to be glycosylated, so if the C-4 hydroxyl is axial,
the monosaccharide is added to the C-3 hydroxyl. However the
products formed in this way do not appear to serve as acceptors
for further transfer reactions. Care must be taken to avoid contam-
ination by wild type enzyme arising either from laboratory proce-
dures or translational misincorporation, since otherwise, initially
formed glycosynthase products can be hydrolyzed, and unexpected
transglycosylation products formed.
acid in ethanol followed by charring. Samples (5
lL) were diluted
in 200 l of 20% MeOH, 0.1% formic acid, and 5
l
lL of the solution
was injected onto an ESI-MS (Waters, Alliance HPLC, Micromass
ZQ ESCi, USA). The quadrupole mass analyzer was scanned over a
range of 100–1000 m/z.
3.4. Transglycosylation by BGlu1 E386G to various acceptors
3. Experimental
3.1. Materials
The BGlu1 E386G glycosynthase was incubated at a concentra-
tion of 40 lM with 10 mM GlcF donor and 10 mM acceptor: glu-
cose, cellobiose, Glc-pNP, Fuc-pNP, Gal-pNP, Man-pNP, or Xyl-
pNP, in 150 mM ammonium bicarbonate, pH 7.0, at 30 °C for
16 h. The products were analyzed by TLC with EtOAc–MeOH–
water (7:2.5:1 as solvent for glucose and pNP-oligosaccharides,
and 2:1:1 for cellobiose). The products were further verified by
Cello-oligosaccharides (DP 3–6) were purchased from Sei-
kagaku Kogyo Co. (Tokyo, Japan). Glucose, cellobiose, Glc2-pNP,
Glc-pNP, Fuc-pNP, Gal-pNP, Man-pNP, and Xyl-pNP were pur-
chased from Sigma Chemical Co. (St. Louis, MO). GlcF, AraF, FucF,
GalF, ManF, and XylF were synthesized by previously described
methods.26 Cyclophellitol was a generous gift from Kah-Yee Li
and Professor Herman Overkleeft, University of Leiden.
ESI-MS, as described above. Samples (3
the reactions with 1:1 molar ratios of GlcF: Glc-pNP or GlcF: Fuc-
pNP were loaded onto ZORBAX carbohydrate column
lL) of the products from
a
(4.6 mm ꢁ 250 mm, Agilent, USA) connected to an Agilent 1100
series LC–MSD. The column was eluted with a linear gradient from
90% to 60% acetonitrile in water over 30 min at a flow rate of
0.8 mL/min. The eluted peaks were detected by UV absorbance at
300 nm. Reaction yields were determined by integration of the
product peaks within the HPLC profile and the molecular masses
of the eluted products were determined by mass spectrometry.
3.2. Protein expression and purification
The recombinant BGlu1 and BGlu1 E386G proteins were ex-
pressed in E. coli strain Origami(DE3) and purified by immobilized
metal affinity chromatography (IMAC), enterokinase digestion, and
subtractive IMAC, as previously described.11,14,27 The buffer was
exchanged with 50 mM phosphate buffer, pH 6.0, and concentrated
in an Amicon ultra-centrifugal filter (Ultracel-10K, Millipore, USA).
The purified protein concentration was determined by measuring
its absorbance at 280 nm. An extinction coefficient e280 of
113560 Mꢀ1 cmꢀ1, calculated from the amino acid composition
by the method of Gill and von Hippel,28 was used to calculate the
BGlu1 enzyme concentrations.
3.5. Determination of linkages within glycosynthase products
Oligosaccharide products from reactions of FucF and Glc2-pNP,
ManF and Glc2-pNP, and GlcF and Fuc-pNP were synthesized with
E386G. Reaction mixtures containing FucF or ManF (14.6 mg,
4 equiv) and Glc2-pNP (9.2 mg, 1 equiv); GlcF (48 mg, 1 equiv)

56
S. Pengthaisong et al. / Carbohydrate Research 352 (2012) 51–59
Figure 4. TLC detection of transglycosylation catalyzed by BGlu1 E386G-2 with various donors or acceptors. The E386G-2 was incubated with 10 mM Glc2-pNP and 10 mM
donors (A and C) or 10 mM GlcF and 10 mM acceptors (B and D) in 150 mM NH4HCO3, pH 7.0, at 30 °C for 16 h. The reactions were performed without (ꢀ) or with (+) enzyme.
The products were separated by silica gel TLC with EtOAc–MeOH–water (7:2.5:1) as a developing solvent and detected under UV light (A and B) or by exposure to 10% sulfuric
acid in ethanol, followed by charring (C and D). G is glucose standard. A and C show reactions with various donors: AF, AraF; FF, FucF; GaF, GalF; GF, GlcF; MF, ManF; and XF,
XylF. B and D show reactions with various acceptors: G, glucose; C2, cellobiose; pG, Glc-pNP; pC, Glc2-pNP; pF, Fuc-pNP; pGa, Gal-pNP; pM, Man-pNP; and pX, Xyl-pNP.
and Fuc-pNP (75 mg, 1 equiv) and 40
l
M of the appropriate en-
(400 MHz) and chemical shifts were reported in d units (ppm)
and were calibrated based on the deuterium solvent chemical shift
(CDCl3). Multiplicity of signals is described as follows: br-broad, s-
singlet; d-doublet; t-triplet; m-multiplet.
zyme were incubated at 30 °C in 150 mM ammonium bicarbonate,
pH 7.0, (2–10 ml total reaction volume) for 24–48 h. The reaction
mixtures were filtered and loaded onto a reverse-phase column
(Sep-Pak tC18, sorbent weight 2 g, Waters), which was eluted with
step gradients of 0–30% methanol in water. Eluted fractions were
checked by TLC, as described above. Fractions containing the com-
pounds were pooled, dried on a rotary evaporator, and further
purified by TLC. Silica gel containing each compound was scratched
from the TLC plate and eluted with 30% MeOH, then the methanol
was evaporated and the sample solution was dried by lyophiliza-
tion. Masses were determined by ESI-MS.
3.5.1. NMR spectra
3.5.1.1. 4-Nitrophenyl (2,3,4-tri-O-acetyl-
a
-
D
-fucopyranosyl)-
1H NMR
(1?3)-2,4-di-O-acetyl-b- -fucopyranoside (1).
D
(400 MHz, CDCl3): d 8.21 (m, 2H, Ar), 7.05 (m, 2H, Ar), 5.50 (dd,
1H, J1,2 8.1, J2,3 9.9 Hz, H-2), 5.35 (br d, 1H, J3,4 3.4 Hz, H-4), 5.20
(br d, 1H, J3 ,4 3.3 Hz, H-40), 5.11 (dd, 1H, J1 ,2 7.9, J2 ,3 10.7 Hz, H-
0
0
0
0
0
0
20), 5.07 (d, 1H, J1,2 8.1 Hz, H-1), 4.95 (dd, 1H, J2 ,3 10.7, J3 ,4
0
0
0
0
0
0
A mixture of pyridine and acetic anhydride (5 ml, Vpyridine:Vacetic
anhydride = 3:2) was added to the purified glycoside products at
room temperature and the reaction mixture was stirred overnight.
Ice water was added to quench the unreacted acetic anhydride and
the reaction was stirred for another 10 min. After evaporation of
the solvent, the residue was dissolved in ethyl acetate and washed
successively with ice water, 1 M HCl, NaHCO3, and brine, and dried
with MgSO4. The organic solvent was evaporated in vacuo, and the
resulting residue was purified by flash column chromatography on
silica gel using ethyl acetate/petroleum ether = 1:1 as eluent.
All 1H and 1H COSY nuclear magnetic resonance (NMR) spectra
3.4 Hz, H-30), 4.55 (d, 1H, J1 ,2 7.9 Hz, H-10), 3.98 (dd, 1H, J2,3 9.9,
J3,4 3.4 Hz, H-3), 3.94 (m, 1H, H-50), 3.76 (m, 1H, H-5), 2.20–1.98
(5s, 15H, 5 ꢁ CH3CO), 1.30 (d, 3H, H60a, H60b, H60c), 1.20 (d, 3H,
H6a, H6b, H6c). ESI-MS: m/z 454 [M+Na]+.
3.5.1.2. 4-Nitrophenyl (2,3,4,6-tetra-O-acetyl-
syl)-(1?3)-2,4-di-O-acetyl-b- -fucopyranoside (2).
a-
D
-glucopyrano-
D
1H NMR
(400 MHz, CDCl3): d 8.21 (m, 2H, Ar), 7.06 (m, 2H, Ar), 5.46 (dd,
1H, J1,2 8.0, J2,3 10.0 Hz, H-2), 5.34 (br d, 1H, J3,4 3.4 Hz, H-4), 5.17
(dd, 1H, J3 ,4 9.3, J4 ,5 9.0 Hz, H-40), 5.12 (dd, 1H, J2 ,3 9.0, J3, 4
0
0
0
0
0
0
0
0
9.3 Hz, H-30), 5.08 (d, 1H, J1,2 8.0 Hz, H-1), 4.93 (dd, 1H, J1 ,2 7.8,
0
0
0
0
0
0
0
0
were recorded on
a
Bruker Avance-400 inv spectrometer
J2 ,3 9.0 Hz, H-20) 4.65 (d, 1H, J1 ,2 7.8 Hz, H-10), 4.33 (dd, 1H, J5 ,6 a

S. Pengthaisong et al. / Carbohydrate Research 352 (2012) 51–59
57
Table 2
Products of glycosynthase reactions of BGlu1 E386G-2 glycosynthase using
a
-glucosyl fluoride donor and different acceptorsa
#
Acceptor
pNP-Oligosaccharide product (% yield)
Di
Tri
Tetra
Penta
Hexa
Hepta
Octa
Nana
—
Total yield
4.2
7.3
(482)
0.040
0.074
(644)
4b
c
4.24
7.37
—
—
—
—
—
—
9.1
11
(498)
0.15
0.21
(660)
5b
c
9.25
11.2
—
—
—
—
—
—
—
—
—
3.0
3.4
(498)
0.49
0.55
(660)
0.15
0.14
(822)
0.12
0.13
(984)
6b
c
3.76
4.22
0.062
0.068
(498)
0.035
0.032
(660)
0.046
0.039
(822)
7b
c
0.143
0.139
—
—
—
—
—
—
—
—
—
—
3.1
4.0
(468)
0.1
0.17
(630)
0.052
0.051
(792)
8b
c
3.25
4.22
26
(660)
16
(822)
11
(984)
3.5
(1146)
0.47
(1308)
0.03
(1471)
0.002
(1633)
9b
57.0
a
Preparative reactions in 150 mM ammonium bicarbonate buffer.
pNP-glycoside acceptor, and 40 lM E386G-2 concentration for 24 h. The relative percents are in terms of peak area per total 300 nm absorbance in pNP-glycoside products
separated by HPLC on a ZORBAX carbohydrate column. The molecular mass [Mass+35Cl]ꢀ of each eluted compound was confirmed by mass spectrometry as shown in
parenthesis.
b
pH 7.0, 30 °C, at a 1:1 molar ratio.
3:1 molar ratios of GlcF donor.
c
2.6, J6 a,6 b 12.3 Hz, H-60a), 4.15 (dd, 1H, J5 ,6 b 3.8, J6 a,6 b 12.3 Hz, H-
60b), 3.97–3.92 (m, 2H, H-3, H-5), 3.69–3.65 (m, 2H, H-50), 2.18–
2.00 (6s, 18H, 6 ꢁ CH3CO), 1.25 (d, 3H, H6a, H6b, H6c). ESI-MS:
m/z 470 [M+Na]+.
1H, J5.6a 1.8, J6a,6b 11.8 Hz, H-6a), 4.52 (d, 1H, J1 ,2 7.9 Hz, H-10),
0
0
0
0
0
0
0
0
4.34 (dd, 1H, J6 a,6 b 12.45, J5 ,6 a 5.2 Hz, H-60a), 4.31–4.28 (m, 2H,
0
0
0
0
H-600a, H-600b), 4.12 (dd, 1H, J6 a,6 b 12.5, J5 ,6 b 2.1 Hz, H-60b), 4.11
(dd, 1H, J6a,6b 11.8, J5,6b 4.9 Hz, H-6b), 3.91–3.81 (m, 3H, H-4, H-
40, H-5), 3.64–3.58 (m, 2H, H-50, H-500), 2.16–1.99 (10s, 30H,
10 ꢁ CH3CO). ESI-MS: m/z 648 [M+Na]+.
0
0
0
0
3.5.1.3. 4-Nitrophenyl [(2,3,4-tri-O-acetyl-b-
(1?4)-O-2,3,6-tri-O-acetyl-b- -glucopyranosyl]-(1?3)-O-2,4-di-
O-acetyl-b- -fucopyranoside (3).
1H NMR (400 MHz, CDCl3):
D-fucopyranosyl)-
D
D
3.5.1.5. 4-Nitrophenyl [(2,3,4-tri-O-acetyl-b-
(1?4)-O-2,3,6-tri-O-acetyl-b- -glucopyranosyl]-(1?4)-O-2,3,6-
tri-O-acetyl-b- -glucopyranoside (5).
1H NMR (400 MHz,
D-fucopyranosyl)-
d 8.21 (m, 2H, Ar), 7.06 (m, 2H, Ar), 5.48 (dd, 1H, J1,2 8.0, J2,3 10.0 Hz,
D
00 00
H-2), 5.33 (br d, 1H, J3,4 3.0 Hz, H-4), 5.20 (br d, 1H, J3 ,4 3.3 Hz, H-
D
400), 5.17–5.03 (m, 4H, H-1,H-30, H-200, H-300), 4.89 (dd, 1H, J1 ,2 7.9,
CDCl3): d 8.21 (m, 2H, Ar), 7.05 (m, 2H, Ar), 5.29 (dd, 1H, J2,3 8.5,
0
0
J2 ,3 8.8 Hz, H-20), 4.58 (d, 1H, J1 ,2 7.9 Hz, H-10), 4.45 (d, 1H, J1 ,2
J3,4 8.2 Hz, H-3), 5.23–5.15 (m, 4H, H-2, H-30, H400, H-1) 5.08 (dd,
0
0
0
0
00 00
8.2 Hz, H-100), 4.30 (dd, 1H, J5,6 a 3.0, J6 a,6 b 12.1 Hz, H-60a), 4.14
1H, J1 2 8.0 J2 3 10.3 Hz, H-200), 4.94 (dd, 1H, J2 3 10.3, J3 4
0
0
0
00 00
00 00
00 00
00 00
(dd, 1H, J5 ,6 b 3.9, J6 a,6 b 12.1 Hz, H-60b), 3.98–3.91 (m, 2H, H-3,
H-5), 3.73–3.63 (m, 3H, H-40, H-50, H-500), 2.19–2.00 (8s, 24H,
8 ꢁ CH3CO), 1.35 (d, 3H, H6a, H6b, H6c), 1.25 (d, 3H, H600a, H600b,
H600c). ESI-MS: m/z 616 [M+Na]+.
3.4 Hz, H-300), 4.87 (dd, 1H, J1 2 8.1, J2 3 9.5 Hz, H-20), 4.55 (m, 1H,
0
0
0
0
0
0
0
0
H-6a), 4.53 (d, 1H, J1 2 8.1 Hz, H-10), 4.43 (d, 1H, J1 2 8.0 Hz, H-
0
0
00 00
100), 4.41 (m, 1H, H-60a), 4.16–4.14 (m, 2H, H-6b, H-60b), 3.88 (dd,
0
0
0
0
1H, J3,4 8.2, J4,5 10.1 Hz, H-4), 3.81 (dd, 1H, J3 4 9.5, J4 5 9.4 Hz, H-
40), 3.91–3.78 (m, 1H, H-5), 3.76–3.71 (m, 1H, H-500), 3.65–3.59
(m, 1H, H-50), 2.20–1.90 (9s, 27H, 9 ꢁ CH3CO), 1.20 (d, 3H, Me).
ESI-MS: m/z 632 [M+Na]+.
3.5.1.4. 4-Nitrophenyl [(2,3,4,6-tetra-O-acetyl-b-
osyl)-(1?4)-O-2,3,6-tri-O-acetyl-b- -glucopyranosyl]-(1?4)-O-
2,3,6-tri-O-acetyl-b- -glucopyranoside (4). NMR
1H
D-mannopyran-
D
D
(400 MHz, CDCl3): d 8.21 (m, 2H, Ar), 7.06 (m, 2H, Ar), 5.40 (br d,
3.6. Generation of E386G-2 mutant of rice BGlu1
1H, J2 ,3 3.0, H-200), 5.29 (dd, 1H, J2,3 8.5, J3,4 8.5 Hz, H-3), 5.21
00 00
00 00
00 00
(dd, 1H, J1,2 7.3, J2,3 8.5 Hz, H-2), 5.20 (dd, 1H, J3 ,4 10.0, J4 ,5
BGlu1 E386G-2 was constructed by changing the DNA codon for
GAA to GGC by QuikChange mutagenesis (Stratagene, La Jolla, CA,
USA) with the primers 50-G ACA GTC GTC ATA ACT GGC AAC GGA
ATG GAT CAA C-30 and 50-G TTG ATC CAT TCC GTT GCC AGT TAT
8.5 Hz, H-400), 5.17 (d, 1H, J1,2 7.3 Hz, H-1), 5.15 (dd, 1H, J2 ,3 9.4,
0
0
J3 ,4 9.7 Hz, H-30), 5.01 (dd, 1H, J2 ,3 3.0, J3 ,4 10.0 Hz, H-300), 4.88
0
0
00 00
00 00
(dd, 1H, J1 ,2 7.9, J2 ,3 9.4 Hz, H-20), 4.64 (br s, 1H, H-100), 4.55 (dd,
0
0
0
0

58
S. Pengthaisong et al. / Carbohydrate Research 352 (2012) 51–59
Table 3
Products of glycosynthase reactions of BGlu1 E386G-2 glycosynthase using different donors and Glc2-pNP acceptora
#
Donor
pNP-Oligosaccharide product (% yield)
Tri
Tetra
Total yield
25
59
(630)
0.051
0.072
(762)
10b
c
25.1
59.1
49
98
(644)
0.011
1.4
(790)
11b
c
49.0
99.4
15
42
(660)
0.013
0.018
(822)
12b
c
15.0
42.0
19
71
(660)
0.21
7.9
(822)
13b
c
19.2
78.9
1.2
3.1
(630)
0.0091
0.013
(762)
14b
c
1.21
3.11
a
Preparative reactions in 150 mM ammonium bicarbonate buffer.
pH 7.0, 30 °C, at a 1:1.
b
c 3:1 molar ratios of glycosyl fluoride donor:Glc2-pNP acceptor, and 40
lM E386G-2 concentration for 24 h. The relative percents are in terms of peak area per total 300 nm
absorbance in pNP-glycoside products separated by HPLC on a ZORBAX carbohydrate column. The molecular mass [Mass+35Cl]ꢀ of each eluted compound was confirmed by
mass spectrometry, as shown by the mass in parenthesis (in amu).
GAC GAC TGT C-30 and the pET32a/BGlu1 expression vector11 as
template. The transglucosylation activity of the new mutation
was compared to the E386G (codon: GGG) toward various donors
of AraF, FucF, GalF, GlcF, ManF, and XylF with Glc2-pNP acceptor,
and GlcF donor with various acceptors of glucose, cellobiose, Glc-
pNP, Glc2-pNP, Fuc-pNP, Gal-pNP, Man-pNP, and Xyl-pNP, as de-
scribed above. Samples (10 ll) of the products from the reactions
with 1:1 and 3:1 molar ratios of donor:pNP acceptor were analyzed
by HPLC on a ZORBAX carbohydrate column as described above.
on galactoside production with BGlu1 E386S glycosynthase. We
also thank Kah-Yee Li and Professor Herman Overkleeft, University
of Leiden, for their generous gift of cyclophellitol. This work was
supported by a Ph.D. scholarship to SP by the Synchrotron Light Re-
search Institute (Public Organization) of Thailand [GS-51-D03],
Suranaree University of Technology National Research University
Project from the Commission on Higher Education of Thailand,
Grant number BRG5380017 from the Thailand Research Fund and
the Natural Sciences and Engineering Research Council of Canada
through Discovery and Strategic Grants and a CREATE training
Grant.
3.7. Cyclophellitol inhibition
The cyclophellitol inhibition of BGlu1 wild type, E386G, and
E386G-2 activities was monitored by analysis of residual enzyme
References
1. Williams, S. J.; Withers, S. G. Carbohydr. Res. 2000, 327, 27–46.
2. Williams, S. J.; Withers, S. G. Aust. J. Chem. 2002, 55, 3–12.
3. Perugino, G.; Trincone, A.; Rossi, M.; Moracci, M. Trends Biotechnol. 2004, 22,
31–37.
4. Fort, S.; Boyer, V.; Greffe, L.; Davies, G. J.; Moroz, O.; Christiansen, L.; Schülein,
M.; Cottaz, S.; Driguez, H. J. Am. Chem. Soc. 2000, 122, 5429–5437.
5. Fairweather, J. K.; Faijes, M.; Driguez, H.; Planas, A. Chem. Biochem. 2002, 3,
866–873.
6. Hrmova, M.; Imai, T.; Rutten, S. J.; Fairweathers, J. K.; Pelosi, L.; Bulone, V.;
Driguez, H.; Fincher, G. B. J. Biol. Chem. 2002, 277, 30102–30111.
7. Kim, Y. W.; Chen, H.; Withers, S. G. Carbohydr. Res. 2005, 279, 42787–42793.
8. Hancock, S. M.; Vaughan, M. D.; Withers, S. G. Curr. Opin. Chem. Biol. 2006, 10,
509–519.
activity as a function of time. Solutions of 0.2
E386G, and 40 M E386G-2 in 50 mM sodium acetate, pH 5.0, were
pre-incubated with 1
lM BGlu1, 40 lM
l
l
M cyclophellitol at 30 °C for different times
and Glc-pNP hydrolysis was assayed as previously described.11
Inactivation of BGlu1 by cyclophellitol was performed by incu-
bating the BGlu1 in 50 mM sodium acetate, pH 5.0, at 30 °C in the
presence of the cyclophellitol concentrations given in Figure 2B.
Aliquots (10 ll) were removed at appropriate time intervals and
assayed for hydrolysis of Glc-pNP in the standard assay.
The b-glucosidase activity of glycosynthase mutants was inacti-
vated by pre-incubating 0.2 lM BGlu1, 40 lM E386G, or 40 lM
9. Faijes, M.; Saura-Valls, M.; Pérez, X.; Conti, M.; Planas, A. Carbohydr. Res. 2006,
341, 2055–2065.
E386G-2 with 2 mM cyclophellitol for 2 h, and then transglycosy-
lation activity was tested with 10 mM Fuc-pNP and 10 mM GlcF
in 150 mM NH4HCO3, pH 7.0, at 30 °C for 24 h. The products were
analyzed by TLC, using EtOAc–MeOH–water (7:2.5:1) as solvent, as
described above.
10. Müllegger, J.; Chen, H.; Chan, W. Y.; Reid, S. P.; Jahn, M.; Warren, R. A. J.; Salleh,
H. M.; Withers, S. G. Chem. Biochem. 2006, 7, 1028–1030.
11. Opassiri, R.; Ketudat Cairns, J. R.; Akiyama, T.; Wara-Aswapati, O.; Svasti, J.;
Esen, A. Plant Sci. 2003, 165, 627–638.
12. Opassiri, R.; Hua, Y.; Wara-Aswapati, O.; Akiyama, T.; Svasti, J.; Esen, A.;
Ketudat Cairns, J. R. Biochem. J. 2004, 379, 125–131.
13. Opassiri, R.; Pomthong, B.; Onkoksoong, T.; Akiyama, T.; Esen, A.; Ketudat
Cairns, J. R. BMC Plant Biol. 2006, 33.
14. Hommalai, G.; Withers, S. G.; Chuenchor, W.; Ketudat Cairns, J. R.; Svasti, J.
Glycobiology 2007, 17, 744–753.
Acknowledgements
15. Chuenchor, W.; Pengthaisong, S.; Robinson, R. C.; Yuvaniyama, J.; Svasti, J.;
Ketudat Cairns, J. R. J. Struct. Biol. 2011, 173, 169–179.
The authors are thankful to Grienggrai Hommalai and Jisnuson
Svasti for initiation of this work and for sharing preliminary data

S. Pengthaisong et al. / Carbohydrate Research 352 (2012) 51–59
59
16. Pengthaisong, S.; Withers, S. G.; Kuaprasert, B.; Svasti, J.; KetudatCairns, J. R.
Protein Sci. 2012, 21, 362–372.
23. Mayer, C.; Zechel, D. L.; Reid, S. P.; Warren, R. A.; Withers, S. G. FEBS Lett. 2000,
466, 40–44.
17. Day, A. G.; Withers, S. G. Biochem. Cell Biol. 1986, 64, 914–922.
18. Parker, J. Microbiol. Rev. 1989, 53, 273–298.
24. Kim, Y. W.; Lee, S. S.; Warren, R. A.; Withers, S. G. J. Biol. Chem. 2004, 279,
42787–42793.
19. Kramer, E. B.; Farabaugh, P. J. RNA 2007, 13, 87–96.
20. Gloster, T. M.; Madsen, R.; Davies, G. J. Org. Biomol. Chem. 2007, 5, 444–446.
21. Withers, S. G.; Umezawa, K. Biochem. Biophys. Res. Commun. 1991, 177, 532–
537.
22. Mackenzie, L. F.; Wang, Q.; Warren, R. A. J.; Withers, S. G. J. Am. Chem. Soc. 1998,
120, 5583–5584.
25. Prade, H.; Mackenzie, L. F.; Withers, S. G. Carbohydr. Res. 1998, 305, 371–381.
26. Yokoyama, M. Carbohydr. Res. 2000, 327, 5–14.
27. Chuenchor, W.; Pengthaisong, S.; Robinson, R. C.; Yuvaniyama, J.; Oonanant,
W.; Bevan, D. R.; Esen, A.; Chen, C. J.; Opassiri, R.; Svasti, J.; Ketudat Cairns, J. R.
J. Mol. Biol. 2008, 377, 1200–1215.
28. Gill, S. C.; von Hippel, P. H. Anal. Biochem. 1989, 182, 319–326.