Enzymatic thioxyloside synthesis: Characterization of thioglycoligase variants identified from A site-saturation mutagenesis library of Bacillus circulans xylanase

Article abstract of DOI:10.1002/cbic.200900711

Thioglycoligases are engineered enzymes for the synthesis of thioglycosides that are derived from retaining glycosidases by replacing the acid/base catalyst. The optimal choice of substitution for the acid/base mutant is currently unknown, so to investigate this question a complete acid/base library of the model glycosidase Bacillus circulans xylanase (Bcx) was generated by using site-saturation mutagenesis. A novel screening approach combining active site titration with semiquantitative product analysis by thin layer chromatography was established and used to evaluate specific activities of each mutant enzyme within crude cell lysates. The six most active Bcx variants were analyzed in more detail, a pH optimum of 8.5 was established and the identity of reaction products was confirmed. Optimal choices for substitution were small, preferably polar amino acids such as threonine, cysteine, and serine. We discuss the resultant data in the context of previously published studies on thioglycoligases.

Full text of DOI:10.1002/cbic.200900711

DOI: 10.1002/cbic.200900711
Enzymatic Thioxyloside Synthesis: Characterization of
Thioglycoligase Variants Identified from A Site-Saturation
Mutagenesis Library of Bacillus Circulans Xylanase
Zachary Armstrong,[a] Stephan Reitinger,[a] Terrence Kantner,[a, b] and Stephen G. Withers*[a]
Thioglycoligases are engineered enzymes for the synthesis of
thioglycosides that are derived from retaining glycosidases by
replacing the acid/base catalyst. The optimal choice of substi-
tution for the acid/base mutant is currently unknown, so to
investigate this question a complete acid/base library of the
model glycosidase Bacillus circulans xylanase (Bcx) was gener-
ated by using site-saturation mutagenesis. A novel screening
approach combining active site titration with semiquantitative
product analysis by thin layer chromatography was established
and used to evaluate specific activities of each mutant enzyme
within crude cell lysates. The six most active Bcx variants were
analyzed in more detail, a pH optimum of 8.5 was established
and the identity of reaction products was confirmed. Optimal
choices for substitution were small, preferably polar amino
acids such as threonine, cysteine, and serine. We discuss the
resultant data in the context of previously published studies
on thioglycoligases.
Introduction
Oligosaccharides play important roles in many biological pro-
cesses,[1–3] thus considerable potential exists for their use as
therapeutics in blocking undesirable processes.[4–6] However,
natural oligosaccharides and glycoconjugates are prone to
degradation by endogenous glycosidases and thus would have
limited utility in such a role. A solution to this problem is the
use of thioglycosides in which the glycosidic oxygen is re-
placed by a sulfur, because such thioglycosides are good
mimics, but are resistant to glycosidase digestion. In recent
years several reports have been published in which mutant
glycosidases, wherein the acid/base residue has been replaced
by a neutral residue (Scheme 1), have been successfully turned
into thioglycoligases, which catalyze the synthesis of thiogly-
cans.[7–14] Incubation of the variant with a donor sugar bearing
a good leaving group such as dinitrophenolate or fluoride re-
sults in the relatively rapid formation of a covalent glycosyl–
enzyme intermediate without need for the acid catalyst. Be-
cause the mutant enzymes also lack the general base catalyst,
the rates of transglycosylation to water or hydroxyl-containing
sugars are extremely low. However, efficient transfer occurs to
acceptor sugars bearing a suitably positioned thiol group, be-
cause the thiolate anion is more nucleophilic and does not
require general base assistance.
ides as enzymatically inert analogues for protein structural
studies on that same system.
Here we report the generation and characterization of a
wide series of acid/base variants of Bcx that function as thio-
glycoligases. Site-saturation mutagenesis was performed to
generate a complete acid/base catalyst mutant library of Bcx.
A novel screening approach in crude cell lysates that combines
the active site titration method[15,16] to quantify native enzyme
with semiquantitative thin layer chromatography (TLC) to
quantify products was used to pick active variants. The best six
of these were characterized in more detail.
Results and Discussion
Generation of acid/base catalyst mutant library and
screening
The enzymatic synthesis of S-glycosidic linkages employing thi-
oglycoligases, a new class of engineered enzymes, is a concept
that has recently emerged. So far four b-glycosidases[7,10–14] and
two a-glycosidases[8] have been successfully turned into thio-
glycoligases by replacing the acid/base catalyst. Activated
While considerable effort has been expended in identifying
optimal replacements for the catalytic nucleophile in glycosyn-
thases by saturation mutagenesis at that position, only one
such study to optimize thioglycoligases has been reported.[12]
Consequently, information on which residues provide opti-
mal substitutions is limited. A well-studied model glycosidase
on which to perform such a study is the Bacillus circulans xyla-
nase Bcx. This has the added advantage that we are interested
in the generation of selective sulfur-linked xylo-oligosacchar-
[a] Z. Armstrong,+ Dr. S. Reitinger,+ T. Kantner, Prof. Dr. S. G. Withers
Department of Chemistry, University of British Columbia
Vancouver BC, V6T 1Z1 (Canada)
Fax: (+1)604-822-2847
E-mail: withers@chem.ubc.ca
[b] T. Kantner
Present address: Pharmacy and Pharmacology, University of Bath
Bath, BA2 7AY (UK)
[+] These authors contributed equally to this work.
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cbic.200900711.
ChemBioChem 2010, 11, 533 – 538
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
533

S. G. Withers et al.
Scheme 1. Mechanism of transglycosylation by a thioglycoligase.
donor sugars are employed to facilitate the formation of the
glycosyl enzyme intermediate in the absence of the acid cata-
lyst. Thiosugar acceptor moieties then accomplish the nucleo-
philic attack without the need for base catalysis, yielding an S-
glycosidic product. By using PCR and the appropriate degener-
ate primers a Bcx gene library containing all 20 possible var-
iants modified at the acid/base position (Glu172) was generat-
ed and cloned into the pET-27b expression system (Novagen).
The mutant library was expressed in Escherichia coli BL21 cells
and the quantities of each active enzyme in the crude cell
lysate and in pure protein preparations were determined by
active site titration employing the mechanism-based inhibitor
2,4-dinitrophenyl 2-deoxy-2-fluoro-b-xylobioside.[16,17] This re-
agent rapidly forms a covalent glycosyl-enzyme intermediate
with stoicheometric release of 2,4-dinitrophenol. Turnover of
this intermediate through hydrolysis is slower than its forma-
tion—leading to a burst phase followed by a steady state. Ex-
amples of such active site titrations for three of the Bcx acid/
base variants are shown in Figure 1. From these curves, the y-
intercept of the back extrapolation of the linear region yielded
the amount of 2,4-dinitrophenol released, thus, through its
extinction coefficient, the enzyme concentration. In the initial
screen the active site titration method affords a very precise
and rapid quantitation of active variants from crude cell lysates
without the need for purification.
To test the relative abilities of each acid/base variant to cata-
lyze the transglycosylation of xylose moieties onto a thio
sugar, mutant enzymes were mixed with donor sugar 2,5-di-
nitrophenyl b-xylobioside (DNP-X2) and acceptor sugar 4-nitro-
phenyl 4-thio b-xylobioside (4S-pNP-X2). Equal amounts of
each enzyme, based upon the active site titration data, were
used and product yield was estimated by semiquantitative TLC
using a digital camera and the software ImageJ (NIH).[19] Mu-
tants in which glutamate 172 was replaced by threonine,
serine, cysteine, valine, alanine, glycine, leucine, isoleucine, me-
thionine, asparagine, glutamine, proline or phenylalanine were
found to produce the expected thioglycoside whereas the
other amino acid replacements, including the wild type
enzyme, did not form detectable amounts of product (Fig-
ure S1 in the Supporting Information). At that stage only the
six most active acid/base variants (threonine, serine, cysteine,
valine, alanine and isoleucine) were chosen for purification and
more detailed analysis by HPLC.
pH profile of Bcx thioglycoligase reaction
Preliminary screening of the alanine variant at pH 6.0, 7.0, and
8.0 indicated that pH 8.0 gave the best yields and tests with
the other variants concurred, with roughly comparable yields
in each case. Consequently, the best of these, the E172T var-
iant, was chosen as a representative to determine the pH-
dependence of the thioglycoligase reaction catalyzed by Bcx
acid/base mutant enzymes. As shown in Figure 2 the largest
amount of product was formed at pH 8.5, with only very little
product formation at pH 6.0, the pH optimum of the native
enzyme for xylan hydrolysis.[18] This corresponds well with pre-
vious studies on another thioglycoligase; these studies showed
that the pH dependence apparently reflects formation of the
thiolate anion as the reactive species.[12] The pKa measured in
this case (~8) is somewhat lower than would be expected for
an alkanethiol, but is likely perturbed downwards due to its
binding in an active site cavity that formerly contained an
anion (Glu172). Interestingly, we observed a decrease in prod-
uct formation at pH values above 8.5. Possible reasons for this
drop in yield at higher pH values include instability of the
Figure 1. Determination of enzyme concentration by active site titration.
The active site titration curves for Bcx E172T, E172S, and E172C are shown
with a linear fit (a) for each. The y-intercept and the extinction coefficient
of 2,4-dinitrophenol were used to calculate the active site concentration.
534
www.chembiochem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2010, 11, 533 – 538

Turning Bacillus circulans Xylanase into a Thioglycoligase
Figure 2. The pH profile for thioglycoside synthesis by Bcx E172T. Relative
yields of product formed from the reaction of DNP-X2 (1 mm) with 4S-pNP-
X2 (1 mm) at a series of pH values. Product was quantitated by HPLC (UV/Vis
detection) relative to an internal standard.
Figure 3. Time course for reaction of the E172T and E172V Bcx variants. Per-
centage yield of product formed from reaction of DNP-X2 (1 mm) with 4S-
pNP-X2 (1 mm) at pH 8.5 at several time points for E172T and E172V. Product
was quantitated by HPLC (UV/Vis detection) relative to an internal standard.
enzyme, greater hydrolysis or possibly a need for a protonated
residue of higher pKa to assist in carboxylic acid departure.
these variants are capable of hydrolyzing the nitrophenyl
moiety from the substrate and/or product–rendering products
“invisible” by UV/Vis. Indeed, TLC analysis of such reaction mix-
tures using an ammonium molybdate/sulfuric acid stain re-
vealed not only the expected pNP-tetrasaccharide product, but
also the free tetrasaccharide and disaccharide. This hydrolytic
reactivity was therefore further probed kinetically.
Characterization of acid/base mutants by HPLC
The efficacies of the six most active variants as thioglycoligases
were analyzed in more detail and the results are summarized
in Table 1. Reactions were performed at the pH optimum
Kinetic studies of the hydrolysis of 4-nitrophenyl
xylobioside
Table 1. Reaction yields for selected Bcx acid/base variants.
Mutant
Yield [%] after 1 h
Yield [%] after 18 h
The kinetic parameters for enzymatic hydrolysis of 4-nitrophen-
yl xylobioside (pNP-X2) catalyzed by the E172S, E172T, E172C,
E172V, E172A and E172I variants in the absence of thiosugar
acceptor are shown in Table 2. The KM value for E172I was sub-
E172T
E172C
E172S
E172V
E172A
E172I
65
62
56
45
43
35
30
47
37
63
58
47
Table 2. Kinetic parameters for the hydrolysis of 4-nitrophenyl xylobio-
side by Bcx variants.
Enzyme
kcat [sÀ1
]
KM [mm]
kcat/KM [sÀ1 mmÀ1
]
(pH 8.5) and aryl xylotetraoside formation was analyzed by re-
versed-phase HPLC with UV/Vis detection. Product was quanti-
fied at two time points by monitoring the absorbance of p-ni-
trophenol (pNP) at 300 nm and quantifying the peak area. At
1 h the most active acid/base mutant enzymes were the threo-
nine and cysteine variants with yields of 65 or 62%, respective-
ly. Interestingly, after 18 h the yields for the variants possessing
polar amino acids (threonine, serine, and cysteine) in place of
the acid/base residue were lower than after 1 h, suggesting
product hydrolysis, while for the nonpolar variants (valine, ala-
nine, and isoleucine) reactions were initially slower, but yields
after 18 h were higher.
Bcx[a]
24
49
0.48
0.014
E172S
E172T
E172C
E172A
E172V
E172I
0.036
0.028
0.015
0.0079
0.0050
n.d.[b]
2.6
3.6
7.0
1.5
6.2
n.d.[b]
0.0076
0.0021
0.0052
0.00082
0.00034
All values are Æca. 10%. [a] Parameters published previously at a pH of
6.0.[22] [b] n.d., not determined.
A more in-depth analysis of time-dependent yield was per-
formed for the E172T and E172V mutant enzymes as represen-
tatives of each class, Figure 3. The threonine variant reacts
faster than the valine variant over the first hour; however, this
is followed by a substantial drop in yield. By contrast the
valine acid/base mutant enzyme starts out more slowly, but
once at its peak the yield decreases only very little, if at all.
The most likely explanation for this behaviour of the threonine
variant, thus likely of the other polar substitutions, was that
stantially higher than the concentration of pNP-X2 that could
be achieved, thus substrate depletion analysis was performed
to determine the kcat/KM value. In general, mutant enzymes
with polar substitutions at the acid/base position exhibit
higher rates of hydrolysis than do those with nonpolar resi-
dues. This corresponds well with the results observed in
Table 1 and Figure 3 where the total amount of pNP-tetrasac-
charide product detected for polar variants is decreased after
longer incubation times.
ChemBioChem 2010, 11, 533 – 538
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
535

S. G. Withers et al.
The only previous such comparison of activities of thioglyco-
ligase variants is that for mutant enzymes developed from
Agrobacterium sp. b-glucosidase (Abg).[12] Interestingly, in that
case substitution of the acid/base catalyst of Abg by a gluta-
mine residue resulted in the highest activity, with the next
highest activity variants being those in which asparagine, gly-
cine, alanine, threonine, and serine were present. Preference
for Gln was particularly great when less activated donor sugars
were employed. A number of other glycosidases (Cellulomonas
fimi b-mannosidase,[10] Escherichia coli a-xylosidase,[8] Sulfolobus
solfataricus a-glucosidase,[8] Xanthomonas manihotis b-galacto-
sidase,[7,13] and Thermotoga maritima b-glucuronidase[11]) have
been successfully converted into thioglycoligases by replace-
ment of the acid/base catalyst with alanine: presumably a
more polar substituent would be a better choice for future de-
velopments. Those results compare well with the findings of
this study. Substitution of the acid/base catalyst with residues
having small, polar or amide side chains appears to work best
to create a thioglycoligase, though care must be taken with
reaction timing to minimize hydrolysis of aryl glycoside bonds
in the products—or noncleavable thioglycosides should be
employed.
Table 3. 1H and 13C NMR chemical shifts and coupling constants of the
thioglycoside product.
d
H-1
H-2
H-3
H-4
H-5a
H-5b
Ring
X1
X2
X3
X4
5.15
4.47
4.39
4.35
3.61
3.21
3.26
3.14
3.55
3.36
3.45
3.31
3.79
2.95
3.69
3.51
4.08
4.01
4.03
3.86
3.52
3.29
3.39
3.19
J Ring
X1
X2
X3
X4
J
1,2 [Hz]
J2,3 [Hz]
9.0
9.0
9.0
9.0
J3,4 [Hz]
9.6
9.6
9.0
9.0
J4,5a [Hz]
8.4
4.8
4.8
5.4
J4,5b [Hz] J5a,5b [Hz]
7.8
9.6
8.4
7.8
10.8
12.0
11.4
12.0
11.4
10.8
12.0
12.0
d
C-1
C-2
C-3
C-4
C-5
Ring
X1
X2
X3
X4
100.2
102.2
84.5
73.8
74.7
72.4
73.1
73.0
72.9
75.4
76.0
76.4
45.6
76.5
69.6
63.5
66.9
66.5
65.6
102.2
Product characterization by mass spectrometry and nuclear
magnetic resonance spectroscopy
rapid method for the evaluation of the library of candidate
thioglycoligases in crude cell extracts. This approach could be
applied much more broadly to speed the analysis of mutant
enzyme libraries. In conjunction with data from previous stud-
ies, we conclude that replacing the acid/base catalyst of a re-
taining glycosidase with a residue having either a small or a
polar side chain or an amino acid with an amide on the side
chain generated the most active thioglycoligases. As discussed
previously[10,12] the inactivity of the wild type enzyme in thio-
glycoside formation is probably a result of deleterious electro-
static interactions rather than steric repulsion, as evidenced by
the activity of the Gln mutants.
Initial confirmation of the formation of the thioglycoside,
rather than an O-glycoside, which would leave a free thiol in
the product, was provided by TLC tests in the presence and
absence of Ellman’s reagent (5, 5’-dithiobis-(2-nitrobenzoic
acid), DTNB). No shift in TLC behaviour was observed. The
structure of the trisaccharide product was fully confirmed by
mass spectrometry and NMR analysis. Indeed the mass spec-
trum (Figure S2) displayed the expected peak at 706 m/z, cor-
responding to the sodium adduct of the product
1
(C26H37NO18S+Na+). Assignments of the 13C- and H NMR spec-
tra (see Table 3 and Figure S3) were carried out by using the
assignments for pNP xylotetraoside as a guide.[19,20] X1 refers to
the xylose residue closest to the reducing end, and X2–X4 are
numbered towards the nonreducing end. The anomeric
carbon on X3 as well as C4 on X2 show substantially lower
chemical shifts than would be expected for an O-linked tetra-
saccharide (d=84.5 for C1X3 and d=45.6 for C4X2 compared
to 102.885 and 77.601 for the O-analogue), thus indicating
that they are attached to a sulfur atom. Spin–spin coupling
constants show that all four anomeric carbons, including the
anomeric centre involved in the sulfur linkage, have the b-con-
figuration (J1,2 =7.8–9.6). These data therefore fully support the
formation of the expected thioglycoside product. Such thiol-di-
rected regioselectivity of transglycosylation reactions catalyzed
by thioglycoligases is consistent with findings on other thiogly-
coligases.[10,11,21]
Experimental Section
Site-saturation mutagenesis and cloning: A PCR approach was
applied to replace the acid/base catalyst E172 with all other possi-
ble amino acid residues, employing the N-terminal primer Bcx_
NdeI and a degenerate primer Rev_AA that binds at the C terminus
(Table S1). 25 PCR cycles were performed. The denaturation step
lasted 30 s at 958C, and the annealing step was performed at 558C
for 40 s. Elongation was performed for 40 s at 728C. The library
was then cloned into pET-27b vector (Novagen) using NdeI and
HindIII restriction enzyme sites and subsequently transformed into
E. coli BL21 DE3 cells. 63 colonies were grown overnight at 378C in
deep-well plates containing LB medium (600 mL) and kanamycin
(50 mgmLÀ1). Sequence analysis revealed that all possible amino
acid substitutions were produced except for E172Q, E172H, and
E172I. In order to complete the acid/base mutant library, the three
missing variants were generated by PCR using individually de-
signed primers (Table S1) and cloned identically into the expression
vector.
Conclusions
Semiquantitative TLC screening combined with an active site
titration method to quantify enzyme provided a novel and
536
www.chembiochem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2010, 11, 533 – 538

Turning Bacillus circulans Xylanase into a Thioglycoligase
Protein expression and purification: Individual clones of E. coli
BL21 DE3 cells, each harbouring a coding sequence of one of the
19 acid/base mutants or the wild type, were grown in deep-well
plates (96 wells) overnight as described above. The next day the
overnight culture (100 mL) was used to inoculate LB medium
(500 mL) containing kanamycin (50 mgmLÀ1) and isopropyl b-d-1-
thiogalactopyranoside (IPTG, 0.7 mm). The cultures were grown
under vigorous shaking at 308C for 4 h, then centrifuged at 3200g
for 15 min (Eppendorf Centrifuge 8510R). Cell pellets were resus-
pended in MES buffer (20 mm, pH 6.0), containing sodium chloride
(50 mm) and polymyxin B sulfate (1 mgmLÀ1) and incubated at
378C for 30 min under gentle shaking. Crude cell extracts obtained
after centrifugation (16000g for 15 min) were used to screen for
thioglycoligase activity by thin layer chromatography (TLC).
ent as a free thiol. Control experiments with the thiol-containing
acceptor sugar confirmed that this assay protocol was effective.
Mass spectrometry and NMR spectroscopy: For detailed charac-
terization of the thioglycoside product a reaction mixture contain-
ing the E172V variant (8.5 mm) and DNP-X2 and 4S-pNP-X2 (2.5 mg
each) in HEPES buffer (20 mm, pH 8.0), NaCl (50 mm), and DTT
(2.5 mm) was incubated overnight at ambient temperature. The re-
action mixture was then filtered (Millex GP 0.22 mm, Millipore) and
loaded onto a reverse-phase cartridge (Sep-Pakꢁ Light tC18, sorb-
ent weight 145 mg, Waters). The column was washed with water
(4 mL) and the product then eluted in 90% methanol. Eluted frac-
tions were further purified by HPLC on a C18 column (particle size
of 5 mm, pore size of 300 ꢂ, Jupiter) using a 0–100% gradient of
acetonitrile in water. Fractions containing the formed thioglycoside
(~2 mg in total) were pooled, dried on a rotary evaporator, then
subsequently dissolved in deuterium oxide (1 mL) for NMR analysis.
The purified product was also analyzed by mass spectrometry
(Waters LC-MS system, ZQ2000).
Pure enzyme preparations for analysis were obtained from 0.5 L
cultures. Cells were induced with IPTG at an OD600 nm ~0.6 (final
IPTG concentration=0.2 mm) and grown at 308C overnight. After
centrifugation cell pellets were resuspended in MES buffer (20 mm,
30 mL, pH 6.0) and treated with lysozyme (25 mg, Sigma Aldrich)
and Benzonaseꢁ Nuclease (125 units, Novagen) at 378C under
gentle shaking for 30 min. After centrifugation (27000g for 15 min)
cell debris was discarded and the supernatant was further diluted
with two volumes of MES buffer and loaded onto SP-Sepharose
ion-exchange columns (HiTrap SP FF, 5 mL, GE Healthcare) and
eluted with a 0–1m NaCl gradient. Fractions containing Bcx var-
iants were pooled and concentrated (Amicon Ultra Centrifugal
Filters Ultracel 10000 NMWL, Millipore) and further purified by
size-exclusion chromatography (HiPrep 16/60 Sephacryl S-100, GE
Healthcare) in HEPES buffer (20 mm, pH 8.0), containing NaCl
(50 mm).
Determination of thioglycoligation reaction yields by HPLC:
Quantification of yields was achieved by analysis of the peak area
corresponding to the eluted thiosugar by using reversed-phase
HPLC (Waters Delta 600E multisolvent delivery system). Reaction
mixtures (200 mL) containing purified enzyme (0.89 mm), NaCl
(50 mm), DNP-X2 (1 mm), 4S-pNP-X2 (1 mm), DTT (2.0 mm), and Tris
(pH 8.5, 20 mm) were incubated at ambient temperature for either
1 or 18 h. A more detailed time course was collected for the E172V
and E172T variants, sampling at several times between 30 min and
18 h. The reactions were terminated by incubation at 858C for
5 min then stored at À708C. The reaction mixture was filtered
(Millex GP 0.22 mm, Millipore) then loaded onto a C18 column (par-
ticle size 5 mm, pore size 300 ꢂ, Jupiter). A gradient of 0–100% ace-
tonitrile in water was used to elute the product and the absorb-
ance at 300 nm was monitored and quantified with the software
Empower (Waters). Para-nitrophenyl cellobioside was employed as
internal standard in all HPLC runs.
Active site titration: To determine the enzyme concentration
crude or purified enzyme preparations (20 mL) were inactivated at
408C in MES buffer (20 mm, 200 mL, pH 6.0), containing NaCl
(50 mm) and the inactivator 2,4-dinitrophenyl 2-deoxy-2-fluoro-b-
xylobioside (2F-DNP-X2, 0.25 mm).[22] The reaction mixture was
placed into a UV/Vis spectrophotometer (UNICAM UV4) equipped
with Vision 3.0 software and the release of 2,4-dinitrophenol was
monitored at a wavelength of 400 nm until a linear rate of hydroly-
sis of the inactivator was observed. The y-intercept of the linear
region was determined by linear regression using the program
GraFit 5.0.[23] The obtained values and the extinction coefficient of
2,4-dinitrophenol (11.40 mmÀ1 cmÀ1) were then used to calculate
the enzyme concentration.
pH profile for Bcx E172T: Yields of the thioglycoligase reaction of
Bcx E172T were determined at pH values ranging from 5.5–9.0. Re-
actions were performed as described above in a series of buffers
(20 mm): MES (pH 5.5–6.5), PIPES (pH 6.5–7.5), and Tris (pH 7.5–9.0).
Overlapping data points from different buffers indicated that no
significant change in activity results from changing buffer identi-
ties. HPLC analysis was achieved as described above.
Kinetic studies of the hydrolysis of 4-nitrophenyl xylobioside:
The rates of enzymatic hydrolysis of pNP-X2 were determined
using a continuous assay. To an appropriate concentration of pNP-
X2 in Tris pH 7.0 (20 mm) and NaCl (50 mm), warmed to 408C, an
aliquot of enzyme was added. Substrate hydrolysis was monitored
by measuring the rate of 4-nitrophenolate release at 400 nm. The
millimolar extinction coefficient of e=9.42 was used to convert
from absorbance units to millimolar values. The kcat/KM value for
E172I was determined by substrate depletion.[25] For this analysis
pNP-X2 (0.0114 mm) in Tris buffer (20 mm), pH 7.0, and NaCl
(50 mm) was incubated at 408C, then an aliquot of enzyme was
added (final concentration of 0.11 mm). The time course for release
of pNP was monitored as above and fit to the equation: A(t)=
Ainf(1 À ekt) +k using GraFit 5.0 software.[23] Division of this rate
constant by the enzyme concentration yielded the kcat/KM value.
Thioglycoligase screening: To identify acid/base variants that cata-
lyze the thioglycoligase reaction, samples of each Bcx variant
(0.1 mm) were assayed in reaction mixtures containing HEPES
(20 mm), NaCl (50 mm), DNP-X2 (donor, 1.6 mm), 4S-pNP-X2 (ac-
ceptor, 1.6 mm), and DTT (2.5 mm) at pH 8.0 at 408C. Aliquots
taken after 10, 15, 20, 25, 30 and 60 min were frozen at À708C
then analyzed by TLC using propan-2-ol/ethyl acetate/water (3:1:1)
as solvent. TLC-plates were visualized under UV-illumination and
the formed product was quantified on digital blots with ImageJ.[24]
Thioglycoside products of the six most active acid/base variants
(threonine, serine, cysteine, valine, alanine and isoleucine) were
also tested for the presence of free thiols by adding an equal
amount (v/v) of a saturated solution of Ellman’s reagent (5,5’-di-
thiobis-(2-nitrobenzoic acid), DTNB) or water to the reaction mix-
ture. Solutions were incubated at room temperature for 5 min and
products were analyzed by TLC. The Rf values of the products did
not shift when pre-incubated with DNTB indicating that the sulfur,
in each case, was involved in an S-glycosidic linkage, thus not pres-
Glossary: Bcx, Bacillus circulans xylanase; DNP-X2, 2,5-dinitrophenyl
b-xylobioside; 4S-pNP-X2, 4-nitrophenyl 4’-thio b-xylobioside; pNP-
X2, 4-nitrophenyl xylobioside; 2F-DNP-X2, 2,4-dinitrophenyl 2-
deoxy-2-fluoro-b-xylobioside; DTT, Dithiothreitol; DNP, dinitrophen-
ChemBioChem 2010, 11, 533 – 538
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
537

S. G. Withers et al.
[9] M. Jahn, H. Chen, J. Mꢃllegger, J. Marles, R. A. Warren, S. G. Withers,
Chem. Commun. 2004, 274–275.
[10] M. Jahn, J. Marles, R. A. Warren, S. G. Withers, Angew. Chem. 2003, 115,
366–368; Angew. Chem. Int. Ed. 2003, 42, 352–354.
[11] J. Mꢃllegger, H. M. Chen, W. Y. Chan, S. P. Reid, M. Jahn, R. A. Warren,
H. M. Salleh, S. G. Withers, ChemBioChem 2006, 7, 1028–1030.
[12] J. Mꢃllegger, M. Jahn, H.-M. Chen, R. A. J. Warren, S. G. Withers, Protein
Eng. Des. Sel. 2005, 18, 33–40.
[13] Y. W. Kim, H. Chen, J. H. Kim, S. G. Withers, FEBS Lett. 2006, 580, 4377–
4381.
yl group; pNP, para-nitrophenyl group; MES, 2-(N-morpholino)etha-
nesulfonic acid; HEPES, (4-(2-hydroxyethyl)-1-piperazineethanesul-
fonic acid); Tris, tris(hydroxymethyl)amino-methane; PIPES, pipera-
zine-N,N’-bis(2-ethanesulfonic acid); DTNB, 5, 5’-dithiobis-(2-nitro-
benzoic acid);
Acknowledgements
[14] M. Jahn, S. G. Withers, Biocatal. Biotransform. 2003, 21, 159–166.
[15] S. Miao, L. Ziser, R. Aebersold, S. G. Withers, Biochemistry 1994, 33,
7027–7032.
[16] S. G. Withers, K. Rupitz, I. P. Street, J. Biol. Chem. 1988, 263, 7929–7932.
[17] L. Ziser, I. Setyawati, S. G. Withers, Carbohydr. Res. 1995, 274, 137–153.
[18] S. L. Lawson, W. W. Wakarchuk, S. G. Withers, Biochemistry 1997, 36,
2257–2265.
[19] E. V. Eneyskaya, H. Brumer, 3rd, L. V. Backinowsky, D. R. Ivanen, A. A.
Kulminskaya, K. A. Shabalin, K. N. Neustroev, Carbohydr. Res. 2003, 338,
313–325.
We are grateful to Johannes Mꢀllegger and Jacqueline Wicki for
fruitful discussions and Hongming Chen for support in synthesiz-
ing xylobioside derivatives. This research was supported by
grants from the Natural Sciences and Engineering Research
Council of Canada (S.G.W.). S.R. gratefully acknowledges the Aus-
trian Science Fund (FWF), for an Erwin Schrçdinger Fellowship.
S.G.W. is the recipient of the Canada Research Chair in Chemical
Biology.
[20] Y. W. Kim, D. T. Fox, O. Hekmat, T. Kantner, L. P. McIntosh, R. A. Warren,
S. G. Withers, Org. Biomol. Chem. 2006, 4, 2025–2032.
[21] R. V. Stick, K. A. Stubbs, Tetrahedron: Asymmetry 2005, 16, 321–335.
[22] S. J. Williams, O. Hekmat, S. G. Withers, ChemBioChem 2006, 7, 116–124.
[23] R. J. Leatherbarrow, GraFit Version 5, Erithacus Software Limited, Horley,
UK, 2001.
Keywords: acid/base catalysts · enzyme catalysis · hydrolases ·
synthesis · thioglycoligases
[1] A. Varki, Glycobiology 1993, 3, 97–130.
[24] W. S. Rasband, ImageJ, U. S. National Institutes of Health, Bethesda,
Maryland, USA http://rsb.info.nih.gov/ij/ 1997–2009.
[25] M. D. Joshi, G. Sidhu, I. Pot, G. D. Brayer, S. G. Withers, L. P. McIntosh, J.
Mol. Biol. 2000, 299, 255–279.
[2] K. Ohtsubo, J. D. Marth, Cell 2006, 126, 855–867.
[3] H. H. Freeze, Nat. Rev. Genet. 2006, 7, 537–551.
[4] S. P. Lockhart, J. G. Hackell, B. Fritzell, Expert Rev. Vaccines 2006, 5, 553–
564.
[5] C. R. Bertozzi, L. L. Kiessling, Science 2001, 291, 2357–2364.
[6] D. B. Werz, P. H. Seeberger, Chem. Eur. J. 2005, 11, 3194–3206.
[7] Y. W. Kim, H. M. Chen, J. H. Kim, J. Mꢃllegger, D. Mahuran, S. G. Withers,
ChemBioChem 2007, 8, 1495–1499.
Received: November 20, 2009
[8] Y. W. Kim, A. L. Lovering, H. Chen, T. Kantner, L. P. McIntosh, N. C. Stry-
nadka, S. G. Withers, J. Am. Chem. Soc. 2006, 128, 2202–2203.
Published online on January 28, 2010
538
www.chembiochem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2010, 11, 533 – 538