A novel -d-galactosynthase from Thermotoga maritima converts -d-galactopyranosyl azide to -galacto-oligosaccharides

Article abstract of DOI:10.1093/glycob/cwq177

The large-scale production of oligosaccharides is a daunting task, hampering the study of the role of glycans in vivo and the testing of the efficacy of novel glycan-based drugs. Glycosynthases, mutated glycosidases that synthesize oligosaccharides in high yields, are becoming important chemo-enzymatic tools for the production of oligosaccharides. However, while β-glycosynthase can be produced with a rather well-established technology, examples of -glycosynthases are thus far limited only to enzymes from glycoside hydrolase 29 (GH29), GH31 and GH95 families. α-l-Fucosynthases from GH29 use convenient glycosyl azide derivatives as a strategic alternative to glycosyl fluoride donors. However, the general applicability of this method to other α-glycosynthases is not trivial and remains to be confirmed. Here, β-d-galactopyranosyl azide was converted to α-galacto- oligosaccharides with good yields and high regioselectivity, catalyzed by a novel -galactosynthase based on the GH36 α-galactosidase from the hyperthermophilic bacterium Thermotoga maritima. These results open a new avenue to the practical synthesis of biologically interesting α-galacto- oligosaccharides and demonstrate more widespread use of α-glycosylα- azide as donors, confirming their utility to expand the repertoire of glycosynthases.

Full text of DOI:10.1093/glycob/cwq177

Glycobiology vol. 21 no. 4 pp. 448–456, 2011
doi:10.1093/glycob/cwq177
Advance Access publication on November 17, 2010
A novel α-D-galactosynthase from Thermotoga maritima
converts β-D-galactopyranosyl azide to
α-galacto-oligosaccharides
Beatrice Cobucci-Ponzano2, Carmela Zorzetti2,
Andrea Strazzulli2, Sara Carillo3, Emiliano Bedini3,
Maria Michela Corsaro3, Donald AComfort4,5, Robert
M Kelly4, Mosè Rossi2,6, and Marco Moracci1,2
Keywords: α-galacto-oligosaccharides / glycosynthase / α-gal
epitope / carbohydrate synthesis / protein engineering
2Institute of Protein Biochemistry – Consiglio Nazionale delle Ricerche, Via
P. Castellino 111, 80131 Naples, Italy; 3Dipartimento di Chimica Organica e
Biochimica, “Università di Napoli Federico II”, Complesso Universitario di
Introduction
Reliable methods for large-scale production of oligosacchar-
ides are currently unavailable, despite the considerable interest
in these molecules. Among other restrictions, this hampers the
development of technical strategies needed to prove their role
in vivo and the efficacy of glycan-based drugs (Brown et al.
2008; Drake et al. 2010; Kiessling and Splain 2010).
Oligosaccharide synthesis is challenging because of the
diverse stereochemistry of the monosaccharide building
blocks, the enormous number of intersugar linkages that they
can form, and the absence, in contrast to nucleic acids and
protein synthesis, of biological processes ruled by universal
conserved codes (Gabius et al. 2004). This challenge is cur-
rently tackled with a variety of approaches, often employed
concurrently, and aimed at automated carbohydrate synthesis
(Seeberger 2008). Classical chemical synthesis, although
powerful, is hampered by laborious regio- and stereochemical
control, while the practical use of highly specific and efficient
glycosyltransferases (GT) is delayed by the costs of substrates
and catalysts (Hancock et al. 2006). Glycoside hydrolases
(GH), working in transglycosylation mode, are an interesting
alternative (McCarter and Withers 1994). These enzymes,
which retain the anomeric configuration of the substrate in the
product, follow a double displacement mechanism, in which a
glycosyl-intermediate is formed and subsequently hydrolyzed
(Figure 1A; Koshland 1953). In the presence of high concen-
trations of acceptors different from water, transglycosylation
occurs. The main drawback of this approach is that the newly
formed product is a substrate of the enzyme and can be
hydrolyzed with reduced final yields. To overcome these
limitations, glycosynthases, a new class of mutant glycosi-
dases in which the active site nucleophile is replaced with a
non-nucleophilic residue, were introduced (Mackenzie et al.
1998; Malet and Planas 1998; Moracci et al. 1998). These
enzymes catalyze the synthesis of oligosaccharides starting
from substrates with inverted anomeric configuration when
compared with the original substrate (Figure 1B, inverting
glycosynthases) (Mackenzie et al. 1998; Malet and Planas
1998), or by exploiting small external nucleophiles
(Figure 1B, retaining glycosynthases) (Moracci et al. 1998).
4
Monte S. Angelo, via Cinthia 4, 80126 Naples, Italy; Department of
Chemical and Biomolecular Engineering, North Carolina State University,
5
EB-1, 911 Partners Way, Raleigh, NC 27695-7905, USA; Department of
Chemical and Materials Engineering, University of Dayton, 300 College Park,
6
Dayton, OH 45469-0246, USA; and Dipartimento di Biologia Strutturale e
Funzionale, Università di Napoli “Federico II”, Complesso Universitario di
Monte S. Angelo, Via Cinthia 4, 80126 Naples, Italy
Received on August 17, 2010; revised on October 25, 2010; accepted on
October 26, 2010
The large-scale production of oligosaccharides is a daunt-
ing task, hampering the study of the role of glycans in
vivo and the testing of the efficacy of novel glycan-based
drugs. Glycosynthases, mutated glycosidases that syn-
thesize oligosaccharides in high yields, are becoming
important chemo-enzymatic tools for the production of
oligosaccharides. However, while β-glycosynthase can be
produced with
a rather well-established technology,
examples of α-glycosynthases are thus far limited only to
enzymes from glycoside hydrolase 29 (GH29), GH31 and
GH95 families. α-L-Fucosynthases from GH29 use con-
venient glycosyl azide derivatives as a strategic alternative
to glycosyl fluoride donors. However, the general applica-
bility of this method to other α-glycosynthases is not
trivial and remains to be confirmed. Here, β-D-galactopyr-
anosyl azide was converted to α-galacto-oligosaccharides
with good yields and high regioselectivity, catalyzed by a
novel α-galactosynthase based on the GH36 α-galactosi-
dase from the hyperthermophilic bacterium Thermotoga
maritima. These results open a new avenue to the practical
synthesis of biologically interesting α-galacto-oligosacchar-
ides and demonstrate more widespread use of β-glycosyl-
azide as donors, confirming their utility to expand the
repertoire of glycosynthases.
1To whom correspondence should be addressed: Tel: +39-0816132271;
Fax: +39-0816132277; e-mail: m.moracci@ibp.cnr.it
© The Author 2010. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com
448

α-Galactosynthase from family GH36
Fig. 1. Reaction mechanism of glycosidases and glycosynthases. (A) Reaction mechanism of a retaining α-galactosidase; R = leaving group; R’OH = H2O:
hydrolysis; R’OH = alcohol: transgalactosylation. (B) Reaction mechanisms of β-glycosynthases; (C) reaction mechanism of α-fucosynthases. (D) Reaction
mechanism of α-galactosynthases.
449

B Cobucci-Ponzano et al.
Several glycosidases from a variety of families of the
Carbohydrate Active enZYme classifications (CAZy) http://
www.cazy.org/ (Cantarel et al. 2009) (namely GH1, GH2,
GH5, GH7, GH8, GH10, GH16, GH17, GH26, GH52 and
GH85 (Perugino et al. 2004; Honda and Kitaoka 2006;
Shaikh and Withers 2008), hydrolyzing β-O-linked sugars
(equatorial), have been converted into efficient glyco-
synthases. In contrast, examples of α-glycosynthases have
thus far been limited to α-L-fucosidases from families GH29
and GH95 (Cobucci-Ponzano et al. 2009; Wada et al. 2008)
and to an α-glucosidase from family GH31 (Okuyama et al.
2002), thereby limiting the access of the glycosynthase
stability of TMGalA. Therefore, these mutants were not
characterized any further.
It was previously reported that Asp327Gly mutant had
its hydrolysis rate and catalytic efficiency reduced by
200–300-fold
on
4-nitrophenyl-α-D-galactopyranoside
(4NP-α-D-Gal) at 37°C if compared with wild-type TmGalA
(Comfort et al. 2007). This inactivation is lower than that com-
monly observed in other GH mutated in the nucleophile
(Viladot et al. 1998; Hrmova et al. 2002; Bravman et al. 2003;
Zechel et al. 2003), but it is not uncommon among
α-glycosidases (Knegtel et al. 1995; Shallom et al. 2002;
Tarling et al. 2003). However, when steady-state kinetic
constants were measured at 65°C, we observed a decrease of
103- and 2.6 × 103-fold of the kcat and kcat/KM, respectively, if
compared with the wild type (Table I). Similar results were
obtained in both sodium acetate and sodium citrate–phosphate
buffers, indicating that acetate ions did not rescue the hydro-
lytic activity of the mutant and that this is the basal activity of
Asp327Gly (Table I).
The chemical rescue of the activity of the Asp327Gly in
the presence of sodium azide was similar to that previously
reported at 37°C (ꢀ30-fold; Table I, Supplementary data,
Figure S1A), and the two enzymes showed a similar depen-
dence on temperature of the α-galactosidase activity
(Supplementary data, Figure S1B; (Comfort et al. 2007).
Possibly, the inactivation produced by the mutation
Asp327Gly was underestimated at 37°C because of the lower
specific activity of the wild type at this temperature.
approach
to
α-fucosylated,
α-galactosylated
and
α-mannosylated oligosaccharides, which are implicated in
important biological phenomena (Vanhooren and Vandamme
1999; Macher and Galili 2008). Along these lines, we recently
reported that two phylogenetically unrelated α-L-fucosidases
could be converted into efficient α-fucosynthases by reactivat-
ing the nucleophile mutants with β-fucosyl-azide (Figure 1C),
as an alternative to the classical β-fluoride derivative donor
(Cobucci-Ponzano et al. 2009). To further confirm that the use
of β-azide derivatives could be of general applicability for
α-glycosynthases, β-galactosyl-azides as possible donors for
α-galactosynthases were considered. To this aim, the
α-galactosidase (TM1192) from the hyperthermophilic bacter-
ium Thermotoga maritima (TmGalA) belonging to family
GH36 was chosen as a model system. Previously, this enzyme
had been characterized and shown to follow the retaining
reaction mechanism in which Asp327 and Asp387 are the
nucleophile and the acid/base, respectively (Comfort et al.
2007) . In addition, the three-dimensional structure at 2.3 Å
resolution has also been deposited (pdb code 1ZY9). We
show here that the mutant Asp327Gly is an efficient
α-galactosynthase producing different galactosylated disac-
charides from β-galactosyl-azide donors and 4-nitrophenyl-α-
and β-glycosides as acceptors. This is the first
α-galactosynthase produced so far and our results demonstrate
that the use of β-azide derivatives is a general strategy for the
preparation of α-glycosynthases.
Asp327Gly uses β-galactosyl-azide as donor for
α-galacto-oligosaccharide synthesis
We have recently reported that two α-L-fucosidases mutated in
the nucleophile were able to promote α-fucosynthetic reac-
tions by using as the donor substrate a fucosyl-azide deriva-
tive with an anomeric configuration opposite to that of the
original substrate (β-fucosyl-azide; Cobucci-Ponzano et al.
2009). By following the same approach, we tested the
TmGalA Asp327Gly mutant for promoting transgalactosyla-
tion reactions from β-galactosyl-azide (β-Gal-N3), which was
promptly chemically synthesized from galactose in three steps
and 85% overall yield (see the Materials and methods
section). Incubations of the mutant at 65°C in 50 mM sodium
acetate buffer, pH 5.0, in the presence of β-Gal-N3 as the
unique substrate (at concentrations 1–14 mM) did not lead to
any product by inspection of runs on thin layer chromato-
graphy (TLC), indicating that no autocondensation reactions
Results and discussion
Kinetic characterization and chemical rescue of TmGalA
nucleophile mutants
The functional role of catalytic nucleophile in the TmGalA has
been assigned to Asp327, based on structural alignment with
other GH36 enzymes, analysis of the pH effect, azide rescue
and characterization of the products obtained with the mutant
Asp327Gly (Comfort et al. 2007). These data prompted us to
test if mutants in Asp327 could act as α-galactosynthases. To
this aim, we analyzed the mutants Asp327Ala and Gly, as
described previously (Comfort et al. 2007), and the newly pre-
pared mutant Asp327Ser. The Asp327Ala and Ser mutant
genes expressed in Escherichia coli were assayed at 65°C in
50 mM sodium acetate, pH 5.0, in permeabilized cells and
were found to be completely inactive as expected. However,
no α-galactosidase activity could be rescued by adding 0.1–2.0
M sodium azide (not shown), suggesting that the two
mutations had a detrimental effect on the activity and/or the
Table I. Steady-state kinetic constants of wild-type and mutant TmGalA
kcat (s−1
)
KM (mM)
kcat/KM (s−1 mM−1
)
Wild type
65
2
0.04 0.02 1585
D327G
0.060 0.003 0.10 0.05
0.060 0.003 0.10 0.04
0.6
0.6
D327Ga
D327G in 500 mM NaN3
2.07 0.04
0.12 0.01
17
Assays were performed in 50 mM sodium acetate, pH 5.0, at 65°C on 2.5
mM 4NP-α-D-Gal.
aAssays were performed under the same conditions but in 50 mM sodium
citrate–phosphate, pH 5.0.
450

α-Galactosynthase from family GH36
were catalyzed under these conditions (not shown). Therefore,
we included different glycosides in the reaction as possible
acceptors at 1:1 and 1:10 donor:acceptor molar ratios. Most of
the acceptors used did not lead to transgalactosylation pro-
ducts under these conditions, suggesting that TmGalA is
rather selective in the −1 subsite of the catalytic center,
according to the current nomenclature (Davies et al. 1997).
However, UV-visible products showing a different migration
compared with substrates were clearly observed by TLC,
when we used 4NP-α-D-Glc (glucopyranoside), -Man (man-
nopyranoside), -Xyl (xylopyranoside) and 4NP-β-D-Xyl as
acceptors at 1:1 molar ratio with the donor (Figure 2,
Supplementary data, Table S1). No transgalactosylation pro-
ducts could be observed when 4NP-α-D-Xyl and 4NP-β-D-Xyl
were used at 1:10 donor:acceptor molar ratios; possibly, under
these conditions, they compete with β-Gal-N3 in the +1 donor
binding site.
To characterize the products, the reaction mixtures were
scaled up and the transglycosylation products were purified, as
described in the Materials and Methods section. In addition,
the transgalactosylation efficiency was measured by running
aliquots of the reaction by High-Performance Anion-Exchange
Chromatography with Pulsed Amperometric Detection
(HPAEC-PAD). These results are summarized in Table II. In
the presence of 4NP-α-Glc, only the compound
1
(α-Gal-(1-6)-α-Glc-4NP) was synthesized in 33% yield (reac-
tion I). Higher yields were obtained when 4NP-α-Xyl and
4NP-β-Xyl were used as acceptors (40 and 38% for reactions
II and III, respectively), obtaining, with the former, compounds
2 and 3 (α-Gal-(1-2)-α-Xyl-4NP and α-Gal-(1-4)-α-Xyl-4NP,
respectively) and only compound 4 (α-Gal-(1-4)-β-Xyl-4NP)
with 4NP-β-Xyl acceptor. It is worth mentioning that the
Asp327Gly mutant showed good regioselectivity with transfer
of the galactose moiety exclusively onto a single OH. When
4Np-α-Glc was the acceptor, the enzyme transgalactosylated
exclusively the primary alcohol at the C6. The other com-
pounds observed by TLC (Supplementary data, Table S1) were
present in negligible amounts and could not be isolated. The
regioselectivity on xyloside acceptors differs on the basis of
the anomeric configuration with main products of transgalacto-
sylation on the OH at the C2 and C4 groups of 4NP-α-Xyl and
4NP-β-Xyl, respectively (Table II). Possibly, this results from
the different binding of these molecules on the −1 subsite. In
reaction IV, with 4NP-α-Man as acceptor, we could isolate two
products, 5 and 6, corresponding to α-Gal-(1-6)-α-Man-4NP
and α-Gal-(1-3)-α-Man-4NP, respectively, with the total yields
of 51%. Again, the primary alcohol at the C6 was the preferred
functional group, whereas the transgalactosylation product on
the OH at the C3 was synthesized at very low amounts as
shown by the relative molar ratio of ꢀ9:1 (Table II). Under
these conditions, much donor substrate was left, but the total
yields, although not quantitative as observed in other glyco-
synthases (Cobucci-Ponzano et al. 2009), are similar to those
of β-glycosynthases from families GH16, GH26, GH52 and
higher than those of the α-fucosynthase from family GH95
(Fairweather et al. 2002; Jahn et al. 2003; Ben-David et al.
2007; Wada et al. 2008). In an effort to increase the efficiency
of the transfer of β-Gal-N3 on 4NP-α-Man (maintaining the
Fig. 2. TLC analysis of the transgalactosylation reactions of the Asp327Gly
mutant. (A) Reactions with 4NP-α-Glc and 4NP-α-Man acceptors. Lane 1,
Gal standard (20 µL of 100 mM); lane 2, β-Gal-N3 standard (20 µL of 80
mM); lane 3, reaction containing Asp327Gly (20 µg), β-Gal-N3 (14 mM),
4NP-α-Glc (14 mM), in 50 mM sodium acetate buffer, pH 5.0, at 65°C for
16 h; lane 4, blank control (same as lane 3, but without enzyme); lane 5,
standard 4NP-α-Glc (20 µL of 100 mM); lane 6, Glc standard (20 µl of 100
mM); lane 7, reaction containing Asp327Gly (20 µg), β-Gal-N3 (14 mM),
4NP-α-Man (14 mM), in 50 mM sodium acetate buffer, pH 5.0, at 65°C for
16 h; lane 8, blank control (same as lane 7, but without enzyme); lane 9,
standard 4NP-α-Man (20 µL of 100 mM); lane 10, Man standard (20 µL of
100 mM). (B) Reactions with 4NP-α-Xyl and 4NP-β-Xyl acceptors. Lane 1,
β-Gal-N3 standard (20 µL of 80 mM); lane 2, reaction containing Asp327Gly
(20 µg), β-Gal-N3 (14 mM), 4NP-α-Xyl (14 mM), in 50 mM sodium acetate
buffer, pH 5.0, at 65°C for 16 h; lane 3, blank control (same as lane 2, but
without enzyme); lane 4, standard 4NP-α-Xyl (20 µL of 100 mM); lane 5,
Xyl standard (20 µL of 100 mM); lane 6, reaction containing Asp327Gly (20
µg), β-Gal-N3 (14 mM), 4NP-β-Xyl (14 mM), in 50 mM sodium acetate
buffer, pH 5.0, at 65°C for 16 h; lane 7, blank control (same as lane 6, but
without enzyme); lane 8, standard 4NP-β-Xyl (20 µL of 100 mM); lane 9,
Gal standard (20 µL of 100 mM). The products were separated on a silica gel
60 F254 TLC using ethyl acetate:methanol:water (70:20:10) as eluent and
were detected by exposure to 4% α-naphtol in 10% sulfuric acid in ethanol
followed by charring. UV-visible reaction products are indicated by ovals.
451

B Cobucci-Ponzano et al.
Table II. Synthetic products of TmGalAD327G
Reaction
I
Donor
Acceptor
Products
Relative ratios (%)
100
II
73
27
III
100
IV
92
8
R = 4-nitrophenol.
452

α-Galactosynthase from family GH36
same 1:1 molar ratio), we used 32-fold more α-galactosynthase
(320 µg in 0.1 mL of reaction mixture): under these con-
ditions, the conversion of the donor increased of 3-fold in a
significantly shortened reaction (1 h), whereas the yields were
1.6-fold lower (data not shown).
To compare our results with the transglycosylating activity
of TmGalA, we incubated the wild type at 65°C in 50 mM
sodium acetate buffer, pH 5.0, 14 mM α-galactosyl-azide and
14 mM 4NP-α-Man acceptor: no products could be observed
(data not shown), demonstrating the utility of the galacto-
synthase approach.
It is worth mentioning that the production of α-galactosides
is generally a challenge in carbohydrate synthetic chemistry,
as with 1,2-cis-glycosides no favorable effect, such as neigh-
boring group participation, is possible (Zhu and Schmidt
2009). Some chemical methods for 1,2-cis-galactosylations
have been reported (Fukase et al. 1998; Demchenko et al.
1999; Imamura et al. 2003; Hada et al. 2006). However, these
are generally based on the use of specific protecting group
patterns, requiring tedious and yield-limiting chemical manip-
ulations for their implementation and typically with a lack of
general utility.
This work represents the proof of principle that the method
can be used for the synthesis of oligosaccharides of techno-
logical interest. Oligosaccharides containing terminal
α-Gal-(1-3)-β-Gal sequence (α-gal epitope), synthesized in
vivo by the α-1,3-galactosyltransferase (α-1,3GT), have been
identified as the major xenoactive antigen responsible for
hyperacute rejection (Galili 2001). In addition, since anti-Gal
is abundant and ubiquitous in humans, the α-gal epitope has
also clinical potential in the production of vaccines and in
anticancer therapies (Macher and Galili 2008). Future work
on the implementation of the α-galactosynthase described
here could open promising alternatives to α-1,3GT for the
production of α-galacto-oligosaccharides.
that β-azide-glycoside derivatives can be usefully exploited
also by GH36 α-galactosynthases, demonstrating that azide
derivatives are a valid alternative to fluoro derivatives for the
production of α-glycosynthases.
Glycosyl-azides are common donors in glycosidases-
catalyzed synthesis and have been used for the screening of
novel glycosidases mutants acting as thioglycoligases
(Müllegger et al. 2005; Bojarová et al. 2007). In our case,
glycosyl azides are crucial for the production of
α-glycosynthases, in fact, β-Fuc-N3 and β-Gal-N3 are more
stable than their β-Fuc-F and β-Gal-F counterparts, because
sodium azide is a less effective leaving group than fluoride.
Therefore, β-glycosyl-azides are not degraded spontaneously
under the operational conditions and, nonetheless, are
sufficiently reactive donors. We demonstrated this property
with the observation that β-fucosyl-, β-galactosyl- and
β-mannosyl-fluorides, being 6-deoxyhexopyranosides and/or
showing axial substituents on the C2 and C4, are easily acti-
vated than hexopyranosides with C4 equatorial substituents,
respectively (Overend 1972; Albert et al. 2000). It is likely
that, the instability of β-fluoride derivatives have hampered the
production of α-glycosynthases; as an attractive alternative,
β-azide derivatives can serve as substrates for the production
of novel α-glycosynthases from unrelated families of
glycosidases.
Materials and methods
Chemicals
All commercially available substrates were purchased from
Sigma-Aldrich (USA). The galactosyl-azides were chemically
synthesized from galactose according to a reported procedure
(Györgydeàk and Szilàgyi 1987): β-Gal-N3 in three steps
(peracetylation, stereoselective anomeric azidation and deace-
tylation) and 85% overall yield while α-Gal-N3 in five steps
(peracetylation, anomeric α-bromination, β-chlorination, azi-
dation and deacetylation) and 16% yield.
Reaction mechanism of α-D-galactosynthase
We demonstrated that the TmGalA mutant Asp327Gly is a
novel α-galactosynthase, the first produced so far, to the best
of our knowledge, and the third example of α-glycosynthases
along with α-glucosynthase and α-fucosynthase (Okuyama
et al. 2002; Wada et al. 2008; Cobucci-Ponzano et al. 2009).
The approach reported here, exploiting β-glycosyl-azides as
donors, is the same recently proposed for two
α-fucosynthases from family GH29 (Cobucci-Ponzano et al.
2009). The mechanism followed by Asp327Gly
α-glycosynthase is shown in Figure 1D: the cavity created by
the mutation that removed the side chain of Asp327 allowed
the access to the active site of the β-Gal-N3. Then, the galac-
tose moiety is transferred to the acceptor activated by the base
catalysis of the Asp387 residue; the disaccharide product,
showing the newly formed α-bond, cannot be hydrolyzed by
the mutant and thus accumulates in the reaction mixture. This
approach follows the same principle proposed for the first
time by Withers and collaborators for β-glycosynthases
exploiting fluoro-glycoside derivatives with an anomeric con-
figuration opposite to that of the “natural” substrate of the
enzyme (Mackenzie et al. 1998). Here, following from GH29
α-fucosynthases (Cobucci-Ponzano et al. 2009), we confirm
Site-directed mutagenesis, expression and purification of
wild-type and mutant TmGalA
The plasmids expressing TmGalA wild type and Asp327Gly
and Ala mutants were described previously (Comfort et al.
2007). The mutant Asp327Ser was generated using the
GeneTailor Site-directed Mutagenesis System (Invitrogen,
USA) and the synthetic oligonucleotides (PRIMM, Italy) in
which the mutated nucleotides are underlined:
TMSerfwd:
5′-GCTACAGGTACTTCAAGATCTCCTTTCTCTTCG-3′
TMSerrev:
5′-GATCTTGAAGTACCTGTAGCCCATCTTTCT-3′
The gene containing the desired mutation was identified by
direct sequencing and completely re-sequenced.
The wild type and mutant α-galactosidases from
T. maritima were expressed in E. coli BL21(DE3) and purified
as described previously (Comfort et al. 2007). After the last
purification step, the enzymes were extensively dialized
against 20 mM sodium phosphate, pH 7.3, and 150 mM
NaCl.
453

B Cobucci-Ponzano et al.
Enzymatic characterization
50% for 15 min for reactions III and IV. For all the disacchar-
ides isolated except for the product 3, the structural determi-
nation was obtained by 1H mono-dimensional and
homonuclear (1H, 1H) and heteronuclear (1H, 13C) two-
dimensional NMR experiments and by methylation analysis.
The α-galactosidase activity assays, the steady-state kinetic
studies and the azide rescue reactions were performed as
reported previously (1 µg of wild type and 10 µg of
Asp327Gly mutant), but at 65°C and without added bovine
serum albumin (Comfort et al. 2007). Suitable blanks, con-
taining all the reagents with the exception of enzyme, were
always used to take into account the negligible spontaneous
hydrolysis of the substrates. One enzymatic unit is defined as
the amount of enzyme catalyzing the conversion of 1 µmol
substrate into product in 1 min, under the indicated con-
ditions. All kinetic data were calculated as the average of at
least two experiments and were plotted and refined with the
program GraFit (Leatherbarrow 1992).
The effect of temperature on the activity was analyzed by
measuring the α-galactosidase activity of wild type and
Asp327Gly mutant (1 and 10 µg, respectively) in 50 mM
sodium acetate buffer, pH 5.0, in the temperature range 37–
80°C. Blanks with no enzyme were used at each temperature
to subtract the spontaneous hydrolysis of the substrate.
Assays of TmGalA wild type and mutants on permeabilized
cells of E. coli BL21(DE3) were performed by the method of
Zhang and Bremer (1995) at 65°C in 50 mM sodium acetate
buffer, pH 5.0, 2.5 mM 4NP-α-Gal and 0.1–2.0 M sodium
azide.
1
Product 1. H NMR (600 MHz, D2O, 308K): δ 5.87 (d, 1H,
JH-1,H-2 = 3.6 Hz, H-1A), 4.88 (d, 1H, JH-1,H-2 = 3.7 Hz,
H-1B), 3.95 (t, 1H, H-3A), 3.93 (d, 1H, H-4B), 3.90 (t, 1H,
H-5A), 3.87 (d,1H, H-6aA), 3.87 (d, 1H, H-5B), 3.82 (dd,
1H, H-2A), 3.75 (d, 1H, H-6aB), 3.74 (d, 1H, H-2B), 3.72 (d,
1H, H-6bB), 3.71 (d, 1H, H-6bA), 3.58 (dd, 1H, H-3B), 3.56
(t, 1H, H-4A); 13C NMR (150 MHz, D2O): δ 99.2 (C-1B),
97.7 (C-1A), 74.5 (C-3A), 73.2 (C-5B), 72.2 (C-2A), 72.2
(C-5A), 70.8 (C-4A), 70.8 (C-3B), 70.5 (C-4B), 69.6 (C-2B),
66.9 (C-6A), 62.4 (C-6B).
1
Product 2. H NMR (600 MHz, D2O, 298K): δ 5.93 (d, 1H,
JH-1,H-2 = 3.3 Hz, H-1A), 5.00 (d, 1H, JH-1,H-2 = 3.8 Hz,
H-1B), 4.09 (t, 1H, H-5B), 3.90 (d, 1H, H-3A), 3.90 (d, 1H,
H-4B), 3.82 (dd,1H, H-3B), 3.79 (dd, 1H, H-2A), 3.69 (t, 1H,
H-5aA), 3.68 (d, 1H, H-2B), 3.67(m, 1H, H-4A), 3.64 (d, 2H,
H-6a,bB), 3.48 (t, 1H, H-5bA); 13C NMR (150 MHz, D2O): δ
97.5 (C-1B), 95.3 (C-1A), 76.0 (C-2A), 72.7 (C-3A), 72.3
(C-5B), 70.3 (C-4B), 70.2 (C-3B), 70.0 (C-4A), 69.3 (C-2B),
63.3 (C-5A), 62.2 (C-6B).
Transgalactosylation trials
1
The glycosynthetic reactions were performed by incubating
Asp327Gly (10 µg) for 16 h at 65°C in 0.1 mL of 50 mM
sodium acetate buffer pH 5.0 at the indicated concentrations
of β-Gal-N3 (donor) and the suitable acceptor. Blank mixtures
without enzyme were always prepared. The products distri-
bution was evaluated by TLC as reported previously
(Cobucci-Ponzano et al. 2009). The transgalactosylation effi-
ciency of Asp327Gly mutant was measured by the use of a
HPAEC-PAD equipped with a PA200 column (Dionex, USA).
Samples were eluted with 20 mM NaOH at a flow rate of 0.5
mL min−1 and the data were analyzed as reported previously
(Cobucci-Ponzano et al. 2009). In particular, to measure the
total amount of galactose enzymatically transferred (total
transgalactosylation efficiency shown in Table II), 1/10 of the
reaction mixtures were incubated for 90 min at 65°C in the
presence of 5.2 µg TmGalA wild type. The efficiency of the
transgalactosylation reaction was calculated as: total amount of
galactose transferred − moles of galactose transferred to water/
total amount of galactose transferred × 100.
Product 4. H NMR (600 MHz, D2O, 298K): δ 5.12 (d, 1H,
J
H-1,H-2 = 7.7 Hz, H-1A), 5.10 (d, 1H, JH-1,H-2 = 3.7 Hz,
H-1B), 4.20 (dd, 1H, H-5aA), 3.87 (dd, 1H, H-3B), 3.86 (t,
1H, H-5B), 3.75 (d, 1H, H-4B), 3.74 (dd,1H, H-2B), 3.69 (t,
1H, H-3A), 3.68 (dd, 1H, H-4A), 3.65 (dd, 2H, H-6a,bB),
3.56 (t, 1H, H-2A), 3.56 (m, 1H, H-5bA); 13C NMR (150
MHz, D2O): δ 101.5 (C-1B), 101.0 (C-1A), 78.7 (C-4A),
75.5 (C-3A), 73.6 (C-2A), 72.7 (C-5B), 70.5 (C-3B), 70.4
(C-4B), 69.7 (C-2B), 65.6 (C-5A), 62.4 (C-6B).
Product 5. 1H NMR (600 MHz, D2O, 315K): δ 5.87 (bs,
1H, H-1A), 4.95 (d, 1H, JH-1,H-2 = 3.8 Hz, H-1B), 4.30 (dd,
1H, H-2A), 4.14 (dd, 1H, H-3A), 4.00 (d, 1H, H-4B), 3.98 (t,
1H, H-5A), 3.97 (dd,1H, H-6aA), 3.90 (t, 1H, H-5B), 3.89 (t,
1H, H-4A), 3.82 (dd, 1H, H-6aB), 3.81 (dd, 1H, H-2B), 3.78
(dd, 1H, H-6bB), 3.74 (dd, 1H, H-6bA), 3.66 (dd, 1H, H-3B);
13C NMR (150 MHz, D2O): δ 99.3 (C-1B), 99.2 (C-1A), 73.8
(C-5B), 72.3 (C-5A), 72.0 (C-3A), 71.1 (C-2A), 71.0 (C-3B),
70.6 (C-4B), 69.8 (C-2B), 68.0 (C-4A), 67.0 (C-6A), 62.4
(C-6B).
Characterization of the galactosylated oligosaccharides
1
The glycosynthetic reactions were performed under the same
conditions described in Transgalactosylation trials section in a
total volume of 2 mL. The transgalactosylation products were
purified by reverse phase chromatography (Polar-RP 80A,
Phenomenex, 4 µm, 250 × 10 mm) on an Agilent HPLC
instrument 1100 series and revealed by UV at 220 nm.
Samples were eluted using the following conditions: 40% of
methanol in water for reaction I; 40% of methanol in water
for 5 min, 40–50% for 30 min, 50% for 15 min for reaction
II; 30% methanol in water for 5 min, 30–50% for 30 min,
Product 6. H NMR (600 MHz, D2O, 298K): δ 5.66 (bs, 1H,
H-1A), 5.23 (d, 1H, JH-1,H-2 = 4.0 Hz, H-1B), 4.28 (dd, 1H,
H-2A), 4.10 (dd, 1H, H-3A), 4.00 (dd, 1H, H-5B), 3.90 (d,
1H, H-4B), 3.85 (dd,1H, H-3B), 3.84 (dd, 1H, H-4A), 3.73
(dd, 1H, H-2B), 3.63 (dd, 1H, H-6aA), 3.65 (dd, 1H, H-6aB),
3.67 (dd, 1H, H-6bB), 3.68 (dd, 1H, H-6bA), 3.56 (dd, 1H,
H-5A); 13C NMR (150 MHz, D2O): δ 101.9 (C-1B), 98.9
(C-1A), 79.3 (C-3A), 75.0 (C-5A), 72.7 (C-5B), 70.7 (C-2A),
70.7 (C-3B), 70.5 (C-4B), 69.5 (C-2B), 66.6 (C-4A), 62.2
(C-6B), 61.7 (C-6A).
454

α-Galactosynthase from family GH36
maritima GH36 α-galactosidase (TmGalA) confirms the mechanistic com-
monality of clan GH-D glycoside hydrolases. Biochemistry. 46:3319–3330.
Davies GJ, Wilson KS, Henrissat B. 1997. Nomenclature for sugar-binding
subsites in glycosyl hydrolases. Biochem J. 321:557–559.
Demchenko AV, Rousson E, Boons GJ. 1999. Stereoselective
1,2-cis-galactosylation assisted by remote neighboring group participation
and solvent effects. Tetrahedron Lett. 40:6523–6526.
Drake PM, Cho W, Li BS, Prakobphol A, Johansen E, Anderson NL, Regnier
FE, Gibson BW, Fisher SJ. 2010. Sweetening the pot: Adding glycosyla-
tion to the biomarker discovery equation. Clin Chem. 56:223–236.
Fairweather JK, Faijes M, Driguez H, Planas A. 2002. Specificity studies of
Bacillus 1,3-1,4-β-glucanases and application to glycosynthase-catalyzed
transglycosylation. ChemBiochem. 3:866–873.
Fukase K, Nakai Y, Kanoh T, Kusumoto S. 1998. Mild but efficient methods
for stereoselective glycosylation with thioglycosides: Activation by
[N-phenylselenophthalimide-Mg(ClO4)2] and [PhIO-Mg(ClO4)2]. Synlett.
1998:84–86.
The disaccharide 3 was isolated in a very low amount pre-
venting us from characterizing it by two-dimensional NMR
experiments. Therefore, the 4NP-α-D-Xyl substitution was
identified by methylation analysis which revealed the presence
of terminal-galactose and 4-substituted xylose, by comparison
1
with authentic standards. The H NMR experiment revealed
for both the xylose (H-1A, δ 5.7) and the galactose units
(H-1B, δ 5.1) an α configuration (JH-1A,H-2A = 3.6 Hz, JH-1B,
H-2B = 3.1 Hz).
The relative ratio reported in Table II was obtained integrat-
ing the UV peaks area in the HPLC chromatogram.
Supplementary data
Gabius HJ, Siebert HC, Andre S, Jimenez-Barbero J, Rudiger H. 2004.
Chemical biology of the sugar code. ChemBiochem. 5:741–764.
Galili U. 2001. The α-gal epitope (Galα1-3Galβ1-4GlcNAc-R) in xenotrans-
plantation. Biochimie. 83:557–563.
Supplementary data for this article is available online at http
://glycob.oxfordjournals.org/.
Györgydeàk ZN, Szilàgyi LS. 1987. Einfache synthesen der anomeren, an
C-6 modifizierten galacto- und glucopyranosylazide. Liebigs Ann.
1987:235–241.
Funding
Hada N, Oka J, Nishiyama A, Takeda T. 2006. Stereoselective synthesis of
1,2-cis galactosides: Synthesis of a glycolipid containing Galα1-6Gal com-
ponent from Zygomycetes species. Tetrahedron Lett. 47:6647–6650.
Hancock SM, Vaughan M, Withers SG. 2006. Engineering of glycosidases
and glycosyltransferases. Curr Opin Chem Biol. 10:509–519.
Honda Y, Kitaoka M. 2006. The first glycosynthase derived from an inverting
glycoside hydrolase. J Biol Chem. 281:1426–1431.
This work was supported by the Agenzia Spaziale Italiana
[project MoMa n. 1/014/06/0]. R.M.K. acknowledges support
from the US National Science Foundation.
Abbreviations
Hrmova M, Imai T, Rutten SJ, Fairweather JK, Pelosi L, Bulone V, Driguez
H, Fincher GB. 2002. Mutated barley (1,3)-β-D-glucan endohydrolases syn-
thesize crystalline (1,3)-β-D-glucans. J Biol Chem. 277:30102–30111.
Imamura A, Ando H, Korogi S, Tanabe G, Muraoka O, Ishida H, Kiso M.
2003. Di-tert-butylsilylene (DTBS) group-directed α-selective galactosyla-
tion unaffected by C-2 participating functionalities. Tetrahedron Lett.
44:6725–6728.
α-1,3GT, α-1,3-galactosyltransferase; Gal, galactopyranoside;
β-Gal-N3, β-galactosyl-azide; GH, glycoside hydrolase;
Glc, glucopyranoside; GT, glycosyltransferases; HPAEC-PAD,
High-Performance
Anion-Exchange
Chromatography
with Pulsed Amperometric Detection; 2- or 4-NP, 2- or
4-nitrophenyl; Man, mannopyranoside; TLC, thin layer chrom-
atography; TmGalA, α-galactosidase from Thermotoga mari-
tima; Xyl, xylopyranoside.
Jahn M, Stoll D, Warren RAJ, Szabo L, Singh P, Gilbert HJ, Ducros VMA,
Davies GJ, Withers SG. 2003. Expansion of the glycosynthase repertoire to
produce
defined
manno-oligosaccharides.
Chem
Commun.
2003:1327–1329.
Kiessling LL, Splain RA. 2010. Chemical approaches to glycobiology. Ann
Rev Biochem. 79:619–653.
Knegtel RMK, Strokopytov B, Penninga D, Faber OG, Rozeboom HJ, Kalk
KH, Dijkhuizen L, Dijkstra BW. 1995. Crystallographic studies of the inter-
action of cyclodextrin glycosyltransferase from Bacillus circulans strain-251
with natural substrates and products. J Biol Chem. 270:29256–29264.
Koshland DE. 1953. Stereochemistry and the mechanism of enzymatic reac-
tions. Biol Rev. 28:416–436.
Leatherbarrow RJ. 1992. GraFit. Staines (UK): Erithacus Software Ltd.
Macher BA, Galili U. 2008. The Galα1,3Galβ1,4GlcNAc-R (α-Gal) epitope:
A carbohydrate of unique evolution and clinical relevance. BBA Gen
Subjects. 1780:75–88.
Mackenzie LF, Wang QP, Warren RAJ, Withers SG. 1998. Glycosynthases:
Mutant glycosidases for oligosaccharide synthesis. J Am Chem Soc.
120:5583–5584.
Malet C, Planas A. 1998. From β-glucanase to β-glucansynthase: Glycosyl
transfer to α-glycosyl fluorides catalyzed by a mutant endoglucanase
lacking its catalytic nucleophile. FEBS Lett. 440:208–212.
McCarter JD, Withers SG. 1994. Mechanisms of enzymatic glycoside
hydrolysis. Curr Opin Struc Biol. 4:885–892.
References
Albert M, Repetschnigg W, Ortner J, Gomes J, Paul BJ, Illaszewicz C,
Weber H, Steiner W, Dax K. 2000. Simultaneous detection of different
glycosidase activities by F-19 NMR spectroscopy. Carbohyd Res.
327:395–400.
Ben-David A, Bravman T, Balazs YS, Czjzek M, Schomburg D, Shoham G,
Shoham Y. 2007. Glycosynthase activity of Geobacillus stearothermophilus
GH52 β-xylosidase: Efficient synthesis of xylooligosaccharides from
α-D-xylopyranosyl fluoride through a conjugated reaction. ChemBiochem.
8:2145–2151.
Bojarová P, Petrásková L, Ferrandi EE, Monti D, Pelantová H, Kuzma M,
Simerská P, Křen V. 2007. Glycosyl azides – An alternative way to disac-
charides. Adv. Synth. Catal. 349:1514–1520.
Bravman T, Belakhov V, Solomon D, Shoham G, Henrissat B, Baasov T,
Shoham Y. 2003. Identification of the catalytic residues in family 52 glyco-
side hydrolase, a β-xylosidase from Geobacillus stearothermophilus T-6.
J Biol Chem. 278:26742–26749.
Brown JR, Crawford BE, Esko JD. 2008. Glycan antagonists and inhibitors:
A fount for drug discovery. Crit Rev Biochem Mol. 42:481–515.
Cantarel BL, Coutinho PM, Rancurel C, Bernard T, Lombard V, Henrissat B.
2009. The Carbohydrate-Active EnZymes database (CAZy): An expert
resource for glycogenomics. Nucleic Acids Res. 37:D233–D238.
Cobucci-Ponzano B, Conte F, Bedini E, Corsaro MM, Parrilli M,
Sulzenbacher G, Lipski A, Dal Piaz F, Lepore L, Rossi M, et al. 2009.
β-Glycosyl azides as substrates for α-glycosynthases: Preparation of effi-
cient α-L-fucosynthases. Chem Biol. 16:1097–1108.
Moracci M, Trincone A, Perugino G, Ciaramella M, Rossi M. 1998.
Restoration of the activity of active-site mutants of the hyperthermophilic
β-glycosidase from Sulfolobus solfataricus: Dependence of the mechanism
on the action of external nucleophiles. Biochemistry. 37:17262–17270.
Müllegger J, Jahn M, Chen HM, Warren RA, Withers SG. 2005. Engineering
of a thioglycoligase: Randomized mutagenesis of the acid-base residue
leads to the identification of improved catalysts. Protein Eng Des Sel.
18:33–40.
Okuyama M, Mori H, Watanabe K, Kimura A, Chiba S. 2002. α-Glucosidase
mutant catalyzes “α-glycosynthase”-type reaction. Biosci Biotech Bioch.
66:928–933.
Comfort DA, Bobrov KS, Ivanen DR, Shabalin KA, Harris JM, Kulminskaya
AA, Brumer H, Kelly RM. 2007. Biochemical analysis of Thermotoga
455

B Cobucci-Ponzano et al.
Overend WG. 1972. The Carbohydrates. London (NY): Academic
Press.
Perugino G, Trincone A, Rossi M, Moracci M. 2004. Oligosaccharide syn-
thesis by glycosynthases. Trends Biotechnol. 22:31–37.
Seeberger PH. 2008. Automated oligosaccharide synthesis. Chem Soc Rev.
37:19–28.
Shaikh FA, Withers SG. 2008. Teaching old enzymes new tricks: Engineering
and evolution of glycosidases and glycosyl transferases for improved gly-
coside synthesis. Biochem Cell Biol. 86:169–177.
Shallom D, Belakhov V, Solomon D, Shoham G, Baasov T, Shoham Y. 2002.
Detailed kinetic analysis and identification of the nucleophile in
Vanhooren PT, Vandamme EJ. 1999. L-fucose: Occurrence, physiological
role, chemical, enzymatic and microbial synthesis. J Chem Technol Biot.
74:479–497.
Viladot JL, de Ramon E, Durany O, Planas A. 1998. Probing the mechanism
of Bacillus 1,3-1,4-β-D-glucan 4-glucanohydrolases by chemical rescue of
inactive mutants at catalytically essential residues. Biochemistry.
37:11332–11342.
Wada J, Honda Y, Nagae M, Kato R, Wakatsuki S, Katayama T, Taniguchi H,
Kumagai H, Kitaoka M, Yamamoto K. 2008. 1,2-α-L-Fucosynthase: A
glycosynthase derived from an inverting α-glycosidase with an unusual
reaction mechanism. FEBS Lett. 582:3739–3743.
α-L-arabinofuranosidase from Geobacillus stearothermophilus T-6,
family 51 glycoside hydrolase. J Biol Chem. 277:51084–51084.
a
Zechel DL, Reid SP, Stoll D, Nashiru O, Warren RAJ, Withers SG. 2003.
Mechanism, mutagenesis, and chemical rescue of β-mannosidase from
Cellulomonas fimi. Biochemistry. 42:7195–7204.
Zhang XG, Bremer H. 1995. Control of the Escherichia coli rrnB P1 promo-
ter strength by ppGpp. J Biol Chem. 270:11181–11189.
Tarling CA, He SM, Sulzenbacher G, Bignon C, Bourne Y, Henrissat B,
Withers SG. 2003. Identification of the catalytic nucleophile of the family
29 α-L-fucosidase from Thermotoga maritima through trapping of a
covalent glycosyl-enzyme intermediate and mutagenesis. J Biol Chem.
278:47394–47399.
Zhu X, Schmidt RR. 2009. New principles for glycoside-bond formation.
Angew Chem Int Ed. 48:1900–1934.
456