β-Glycosyl Azides as Substrates for α-Glycosynthases: Preparation of Efficient α-L-Fucosynthases

Article abstract of DOI:10.1016/j.chembiol.2009.09.013

Fucose-containing oligosaccharides play a central role in physio-pathological events, and fucosylated oligosaccharides have interesting potential applications in biomedicine. No methods for the large-scale production of oligosaccharides are currently avai

Full text of DOI:10.1016/j.chembiol.2009.09.013

Chemistry & Biology
Article
b-Glycosyl Azides as Substrates
for a-Glycosynthases: Preparation
of Efficient a-L-Fucosynthases
Beatrice Cobucci-Ponzano,1 Fiorella Conte,1 Emiliano Bedini,2 Maria Michela Corsaro,2 Michelangelo Parrilli,2
3
3
4
4
Gerlind Sulzenbacher, Alexandra Lipski, Fabrizio Dal Piaz, Laura Lepore, Mose Rossi,1,5 and Marco Moracci1,
*
`
1Institute of Protein Biochemistry – Consiglio Nazionale delle Ricerche, Via P. Castellino 111, 80131, Naples, Italy
`
Dipartimento di Chimica Organica e Biochimica, Universita di Napoli Federico II, Complesso Universitario di Monte S. Angelo,
2
via Cinthia 4, 80126 Napoli, Italy
3
`
`
`
`
`
Architecture et Fonction des Macromolecules Biologiques, UMR6098 CNRS, Universite de Provence, Universite de la Mediterranee,
Case 932, 163 Avenue de Luminy, 13288 Marseille Cedex 9, France
4
`
Dipartimento di Scienze Farmaceutiche, Universita di Salerno, Via Ponte Don Melillo, 84084 Fisciano, Salerno, Italy
Dipartimento di Biologia Strutturale e Funzionale, Universita di Napoli Federico II, Complesso Universitario di Monte S. Angelo,
5
`
via Cinthia 4, 80126 Napoli, Italy
*Correspondence: m.moracci@ibp.cnr.it
DOI 10.1016/j.chembiol.2009.09.013
SUMMARY
from the laborious regio- and stereochemical control. An inter-
esting option is enzymatic synthesis promoted by glycosyltrans-
Fucose-containingoligosaccharides play a central role ferases (GT) and glycoside hydrolases (GH) working in transgly-
cosylation mode. The former enzymes are extremely efficient,
in physio-pathological events, and fucosylated oligo-
but their use is hampered by the high cost of the sugar nucleotide
substrates, the difficulties in obtaining the catalysts in sufficient
saccharides have interesting potential applications in
biomedicine. No methods for the large-scale produc-
tion of oligosaccharides are currently available, but
thechemo-enzymatic approach is very promising. Gly-
cosynthases, mutated glycosidases that synthesize
oligosaccharides in high yields, have been demon-
amounts, and the extreme specificity of the enzymes (Hancock
et al., 2006). Recent successes in improving the promiscuity of
GTs by directed evolution have opened new perspectives in
the use of these enzymes, but their exploitation is still limited
(Aharoni et al., 2006; Williams et al., 2007). An alternative to
strated to be an interesting alternative. However,
examples of glycosynthases available so far are
restricted to a limited number of glycosidases families
glycosyltransferases, GHs that retain the anomeric configuration
in the product, are widespread enzymes that, using cheap
substrates, promote oligosaccharide synthesis by transglycosyla-
and to only one retaining a-glycosynthase. We show tion reactions. Retaining glycosidases utilize a double displace-
ment mechanism in which a glycosyl intermediate is formed and
subsequently hydrolyzed (see Figure S1 available online): when
acceptorsdifferentfromwaterintercepttheglycosyl-enzymeinter-
here that new mutants of two a-L-fucosidases are effi-
cient a-L-fucosynthases. The approach shown utilized
b-L-fucopyranosyl azide as donor substrate leading to
transglycosylation yields up to 91%. This is the first
method exploiting a b-glycosyl azide donor for a-gly-
cosynthases; its applicability to the glycosynthetic
methodology in a wider perspective is presented.
mediate, transglycosylation reactions occur (McCarter and
Withers, 1994). This approach might allow regio- and stereospec-
ificity control, but yields are usually not higher than 40% because
the products of the reaction are, at the same time, substrates. To
overcome these drawbacks, glycosynthases, a new class of
mutant glycosidases derived from exo- and endo-b-glucosidases
INTRODUCTION
by replacing the active site nucleophile with a nonnucleophilic
residue, were introduced (Mackenzie et al., 1998; Moracci et al.,
The great structural variety of carbohydrates results from the 1998; Malet and Planas, 1998). The mutation completely inacti-
diverse stereochemistry of the monosaccharide building blocks vates the enzyme, but, in the presence of a substrate with good
and from the enormous number of intersugar linkages that they leaving group ability, the activity of the mutant can be restored.
can form. This feature makes sugar molecules suitable for medi- In fact, the small cavity created upon mutation can accommodate
ating many biological processes (Varki, 1993), but it complicates a substrate with inverted anomeric configuration when compared
greatly their production. In addition, carbohydrate synthesis withthe original substrate (Figure S2A, inverting b-glycosynthases)
in vivo, in contrast to nucleic acids and protein synthesis, is (Mackenzie et al., 1998; Malet and Planas, 1998), or a small anion
a very complex process that is not regulated by universally (Figure S2B, retaining b-glycosynthases) (Moracci et al., 1998). In
conserved codes, and has not been automated in vitro yet. both cases, the mutant enzyme promotes the transglycosylation
This challenge currently motivates the efforts in developing to an acceptor with almost quantitative yields. Inverting b-glyco-
methods for large-scale production of oligosaccharides (See- synthases cannot hydrolyze the product of the reaction showing
berger, 2008). At present, no such methods are available ab-anomericconfiguration, whereasinretaining b-glycosynthases
because, in chemical synthesis, most of the difficulties arise the bad leaving ability of the glycosidic group in the products
Chemistry & Biology 16, 1097–1108, October 30, 2009 ª2009 Elsevier Ltd All rights reserved 1097

Chemistry & Biology
a-L-Fucosynthases
prevents their subsequent hydrolysis, resulting in quantitative SsD242G (Cobucci-Ponzano et al., 2003b). As expected, both
yields.
mutants were almost completely inactive, but the hydrolytic
The approach leading to inverting glycosynthases is suitable activity of SsD242S was chemically rescued with external nucle-
for both exo- and endo-glycosidases and has successfully ophiles resulting in increased activity on both 4-nitrophenylꢀ and
been applied to a variety of GHs belonging to several families 2-chloro-4-nitrophenyl a-L-fucopyranosides (4NP-a-L-Fuc and
of the carbohydrate active enzyme classification (Cantarel 2C4NP-a-L-Fuc) (Figure S3) (for a review on the chemical rescue
et al., 2009; http://www.cazy.org/), namely, GH1, GH2, GH5, of the activity of glycoside hydrolase mutants see Ly and
GH7, GH10, GH16, GH17, GH26, GH31, and GH52. Instead, Withers, 1999). The increased activity in sodium formate was
retaining glycosynthases have been obtained only from hyper- similar for both substrates (Figure S3A), whereas in sodium
thermophilic exo-b-glycosidases from family GH1 (for reviews azide, SsD242S showed an about 3-fold higher specific activity
see Perugino et al., 2004, 2005, and Hancock et al., 2006).
on 2C4NP-a-L-Fuc if compared with 4NP-a-L-Fuc (Figure S3B).
In spite of the convenience of this approach, GHs recalcitrant to In our hands, SsD242A was not reactivated by sodium azide and
become glycosynthases are not uncommon (Ducros et al., 2003; was not further characterized.
Cobucci-Ponzano et al., 2003b; Perugino et al., 2005), which
The steady-state kinetic parameters of the mutants on the
explains why efforts in developing newmethods to improve glyco- 4NP-a-L-Fuc and 2C4NP-a-L-Fuc substrates were measured
synthases are continuous (Kim et al., 2004; Lin et al., 2004; Ben- in several conditions (Table 1); remarkably, the turnover number
David et al., 2008). Recently, glycosynthases have been prepared of SsD242S on 2C4NP-a-L-Fuc in sodium azide (2 M) was 1.8-
from enzymesfollowingatypicalreaction mechanisms(GH85)and fold higher than that of the wild-type.
from inverting glycosidases (GH8 and GH95) (Umekawa et al.,
2008; Honda and Kitaoka, 2006; Honda et al., 2008; Wada et al., Analysis of the Products of Ssa-fuc Mutants
2008). However, these approaches (as stated by the authors), Incubation of SsD242G and SsD242S in the presence of 4NP-a-
though innovative, do not provide common strategies to convert L-Fuc or 2C4NP-a-L-Fuc (2 and 20 mM, respectively) and
an inverting glycosidase into a glycosynthase (Wada et al., 2008). sodium formate did not lead to observable transfucosylation
The most noticeable examples of enzymes that are not prone products on thin-layer chromatography (TLC), confirming the
to function as glycosynthases are retaining a-glycosidases with results previously reported (Cobucci-Ponzano et al., 2008).
only a GH31 a-glucosidase available to date (Okuyama et al., Instead, newly formed transfucosylation products were identi-
2002). The scarcity of a-glycosynthases has hampered so far fied in the presence of sodium azide (Figure 1). In the presence
the access to the synthesis of a large class of oligosaccharides of 2C4NP-a-L-Fuc the mutants produced two compounds,
of biotechnological interest, such as a-L-fucosylated oligosac- which were not UV-visible on TLC. Remarkably, after incubation
charides. Fucose-containing oligosaccharides play a central with the wild-type Ssa-fuc, one of the two compounds was
role in a number of physiological and pathological events (Ma completely hydrolyzed, demonstrating that it contained a-L-
et al., 2006), and therefore fucosylated oligosaccharides have anomeric bonds (Figure 1B). The combined use of nuclear
potential applications in biomedicine (Vanhooren and Van- magnetic resonance and mass spectrometry allowed us to
damme, 1999). However, the synthesis of a-L-fucosides by clas- unequivocally define the structure of products 1 and 2 (Fig-
sical chemical methods is challenging because of the remarkable ure 1C). Compound 1, b-L-fucopyranosyl azide (b-L-Fuc-N3),
instability of fucosyl donors commonly used in chemical glycosyl- which is produced with both 4NP- and 2C4NP-a-L-Fuc, is
ations, explaining the interest in enzymatic approaches. a-L- not surprising: SsD242G produced b-L-Fuc-N3 in similar condi-
Fucosidase-catalyzed reactions led to transfucosylating yields tions as expected from a mutant in the nucleophile (Cobucci-
in the range 6%–54% (Murata et al., 1999; Farkas et al., 2000; Ponzano et al., 2003b). Instead, compound 2, the disaccharide
Wada et al., 2008), whereas mutants obtained by directed evolu- a-L-fucopyranosylꢀ(1-3)-b-L-fucopyranosyl azide (a-L-Fuc-
tion of the a-L-fucosidase from Thermotoga maritima (Tma-fuc) (1-3)-b-L-Fuc-N3), is unexpected. It is produced only from the
increased the transfucosylation yields from 7% (wild-type) to 2C4NP-a-L-Fuc and more efficiently by SsD242S (Figure 1B);
60% (Osanjo et al., 2007). Noticeably, this result was obtained presumably, the better leaving ability of the 2-chloro-4-nitrophenol
by mutating three amino acids not directly involved in catalysis.
aglycon, if compared with the 4-nitrophenol, and the mutation in
In the past, we have tried to convert by mutation the retaining Ser were decisive to improve the transfucosylation reaction. The
a-L-fucosidase from the hyperthermophilic archaeon Sulfolobus presence of the b-L-Fuc-N3 group at the reducing end of this
solfataricus(Ssa-fuc)intoanovela-L-fucosynthaseunderavariety compound (Figure 1C) indicated that no transfucosylation
of conditions, but no oligosaccharide product was observed (Co- occurred on 2C4NP-a-L-Fuc, suggesting that it was an efficient
bucci-Ponzano et al., 2008). Here we describe a novel strategy for donor, but a poor acceptor.
the production of efficient fucosynthases by using b-L-fucopyra-
nosyl azide as donor substrate. The applicability to the glycosyn- Characterization of the Reaction Mechanism
thetic methodology in a wider perspective is also presented.
of the Ssa-fuc Mutants
The most surprising result of the sodium azide activity rescue
experiments is that mutants in the nucleophile of Ssa-fuc
promoted the formation of both b-L- and a-L-bonds. Two alter-
native reaction mechanisms could explain how the Ssa-fuc
mutants promoted the synthesis of compound 2 containing an
RESULTS
Construction of Ssa-fuc Nucleophile Mutants
and Kinetic Characterization
The SsD242A and SsD242S mutants of the a-L-fucosidase from a-L-bond. In one case, the b-L-Fuc-N3 product might become
S. solfataricus were produced as previously described for a novel donor and acceptor as a result of its accumulation;
1098 Chemistry & Biology 16, 1097–1108, October 30, 2009 ª2009 Elsevier Ltd All rights reserved

Chemistry & Biology
a-L-Fucosynthases
sodium azide and the hypothetical nucleophile compete in the
attack to the anomeric center of 2C4NP-a-L-Fuc (in Figure 2B
compare the upper and lower schemes). To verify this latter
hypothesis, we characterized single and double mutants in the
residues Glu58 and Glu292, which cooperate with Asp242 in
catalysis (Cobucci-Ponzano et al., 2005), but we could not find
evidence of their involvement as novel nucleophile. In addition,
assuming that a residue acting as nucleophile in SsD242S might
react with a mechanism-based inhibitor, we prepared the inhib-
itor 2-deoxy-2-fluoro-a-L-fucosyl fluoride (2d-2F-Fuc-F, see
Supplemental Data), aiming to identify this hypothetical residue
by high-performance liquid chromatography (HPLC)/electro-
spray mass spectrometry as previously reported (Tarling et al.,
2003). The mechanism-based inhibitor bound to Asp242 in the
wild-type Ssa-fuc used as control whereas the treatment of
SsD242S did not produce any labeled peptide (see Supple-
mental Data). Though these negative results are not conclusive,
our data strongly indicate that no additional catalytic nucleophile
was produced by mutating Asp242 in Ssa-fuc, making question-
able the validity of the reaction mechanism shown in Figure 2B.
To validate the reaction mechanism described in Figure 2A, we
tested if SsD242S was able to use b-L-Fuc-N3 as donor.
Remarkably, SsD242S incubated for 16 hr at 65ꢁC in 20 mM
b-L-Fuc-N3 catalyzed the formation of the disaccharide 2 with
a transfucosylation efficiency, defined as the amount of fucose
transferred to an acceptor different from water, measured by
HPAEC-PAD, of 40% and exclusive formation of the a-L-(1-3)-
bond (Figure 3A). Similar results were obtained with SsD242G,
but TLC inspection indicated lower efficiency. b-L-Fuc-N3 re-
mained unreacted in the absence of the mutant (Figure 3A) or
in the presence of the wild-type Ssa-fuc incubated at the same
conditions (data not shown), ruling out possible artifacts.
These results confirm the reaction mechanism described in
Figure 2A in which the b-L-Fuc-N3, produced by the mutants
from the 2C4NP-a-L-Fuc and azide, become a donor during
the course of the reaction.
Table 1. Steady-State Kinetic Constants of Wild-Type and Mutant
a-L-Fucosidases
kcat
(sꢀ1
KM
kcat/KM
(sꢀ1 mMꢀ1
)
)
(mM)
Wild-type Ssa-fuca
4NP-a-L-Fuc
287 11 0.028 0.004 10250
2C4NP-a-L-Fuc
SsD242G
157
9
0.013 0.004 11602
4NP-a-L-Fuc
0.24b
—
—
+ sodium formate pH 4.0
4NP-a-L-Fuc
5.9 0.2 1.0 0.1
1.6 0.1 0.2 0.1
6
7
2C4NP-a-L-Fuc
+ sodium azide 2 M
4NP-a-L-Fuc
9.7 0.3 0.19 0.02
51
2C4NP-a-L-Fuc
SsD242S
55
3
0.14 0.02
384
4NP-a-L-Fuc
0.08b
0.25b
—
—
—
—
2C4NP-a-L-Fuc
+ sodium formate
4NP-a-L-Fucc
3.6 0.1 0.9 0.1
4
2C4NP-a-L-Fucd
+ sodium azide 2 M
4NP-a-L-Fuc
5.8 0.2 0.045 0.005 129
47
1
0.19 0.02
247
987
2C4NP-a-L-Fuc
Wild-type Tma-fuce
4NP-a-L-Fuc
286 35 0.3 0.1
80
88
3
5
0.033 0.005 2412
0.007 0.003 12287
2C4NP-a-L-Fuc
TmD224G
4NP-a-L-Fuc
0.05b
0.46b
—
—
—
—
2C4NP-a-L-Fuc
+ sodium azide 1 M
4NP-a-L-Fuc
9.2 0.3 0.015 0.002 613
27 0.040 0.005 681
2C4NP-a-L-Fuc
1
a Assays were performed in 50 mM sodium phosphate buffer (pH 6.5)
at 65ꢁC.
Construction and Characterization of Tma-fuc
Nucleophile Mutants
b kcat was determined from the initial velocity at saturating concentration
The use of the b-L-fucosyl azide donor could be a novel general
strategy for the production of efficient fucosynthases, therefore,
we prepared mutants in Gly and Ser of the catalytic nucleophile
Asp224 of Tma-fuc, which shows rather low (25%) amino acid
sequence identity to Ssa-fuc, but has been fully characterized
(Tarling et al., 2003; Sulzenbacher et al., 2004).
of substrate.
c In 1 M sodium formate.
d In 2 M sodium formate.
e Assays were performed in 50 mM sodium citrate/phosphate buffer
(pH 6.0) at 60ꢁC. The standard deviation is reported.
As expected, mutations severely affected the fucosidase
activity, which could not be rescued in the TmD224S with both
thus, compound 1 is transferred to another b-L-Fuc-N3 molecule sodium azide or sodium formate; thus, this mutant was not
leading to a-L-Fuc-(1-3)-b-L-Fuc-N3 (Figure 2A, lower diagram). further analyzed. Instead, the residual activity of the TmD224G
This mechanism would imply that azide in b-L-Fuc-N3 could mutant (0.05 and 0.46 sꢀ1 on 4NP-a-L-Fuc and 2C4NP-a-L-
work as leaving group; though not impossible, this is certainly Fuc, respectively) was rescued only by sodium azide on both
surprising because glycosyl-azide compounds are rather stable substrates (Figure S4). The specific activity was higher on
(Ly and Withers, 1999). Therefore, an alternative explanation 2C4NP-a-L-Fuc, as reported above for SsD242S (compare
could be that a third amino acid in the active site of the enzyme Figures S3 and S4).
might have gained the function of nucleophile as a consequence
The steady-state kinetic constants of TmD224G in sodium
of the mutation of the Asp242. This hypothetical residue might azide are compared with those of the wild-type in Table 1. The
form the covalent intermediate and then transfer the fucose to specificity constants of the mutant were 4- and 18-fold lower
the b-L-Fuc-N3 acceptor (Figure 2B, lower diagram). At the than those of the wild-type on 4NP-a-L-Fuc and 2C4NP-a-L-
earliest stages of the reaction, following this second hypothesis, Fuc, respectively. However, the affinity for both substrates was
Chemistry & Biology 16, 1097–1108, October 30, 2009 ª2009 Elsevier Ltd All rights reserved 1099

Chemistry & Biology
a-L-Fucosynthases
Figure 1. TLC Detection of the Transfuco-
sylation Products of SsD242G/S in Sodium
Azide
(A) 2 M sodium azide and 2 mM 4NP-a-L-Fuc; lane
1: fucose marker; lane 2: SsD242G reaction; lane 3:
SsD242S reaction; lane 4: blank with no enzyme.
(B) 0.1 M sodium azide and 20 mM 2C4NP-a-L-
Fuc: lane 1: SsD242G reaction; lane 2: SsD242S
reaction; lanes 3 and 4: the same samples shown
in lanes 1 and 2, respectively, after incubation in
the presence of wild-type Ssa-fuc; lane 5: blank
with no enzyme.
(C) Based on later product analysis, 1 and 2 turned
out to be b-L-Fuc-azide and a-L-Fuc-(1,3)-b-L-
Fuc-azide, respectively.
These results showed that the mutants
have wide specificity for the acceptor
molecule. It is worth noting that most of
the donor was converted and all the
products accumulated in the reaction
after prolonged incubation, as expected
for efficient fucosynthases.
To determine the regioselectivity and
the ability of SsD242S and TmD224G in
synthesizing longer oligosaccharides,
we analyzed preparative reactions of the
two enzymes. Table 2 summarizes the
results of these synthetic trials: reactions
not significantly impaired by the mutation, remaining in the I and II show the products of SsD242S in 10 mM of b-L-Fuc-N3
micromolar range.
donor with 4NP-b-D-Xyl and 4NP-b-D-Gal acceptors (34 mM
TLC analysis of the reaction mixtures of TmD224G after incu- and 20 mM, respectively), whereas in reaction III the mutant
bation at 60ꢁC for 16 hr in sodium azide 0.05–1.0 M and 20 mM was incubated with 5 mM donor and 15 mM 4NP-b-D-GlcNAc
2C4NP-a-L-Fuc revealed a single transfucosylation product of acceptor. In reaction IV, TmD224G mutant was incubated
the same polarity of b-L-Fuc-N3 while no disaccharide products with 10 mM b-L-Fuc-N3 as donor and 100 mM 4NP-b-D-Xyl
were formed (data not shown). However, incubations in the pres- acceptor.
ence of 10 mM b-L-Fuc-N3 and 4NP-b-D-Xyl at 1:5 and 1:10
The products of the reaction were isolated by reverse-phase
donor:acceptor molar ratios revealed one UV-visible transfuco- HPLC and each product was identified by matrix-assisted laser
sylation product, showing that TmD224G exploits b-L-Fuc-N3 desorption/ionization time-of-flight (MALDI-TOF) mass spectrom-
for its a-fucosynthase activity (Figure 3B).
etry, methylation analysis, and 1H-NMRspectroscopy. Inthereac-
tions containing SsD242S we always found compound 2 and, in
Oligosaccharide Synthesis by a-L-Fucosynthases
trace amounts, a-L-Fuc-N3 ([M+Na]+ = m/z 212.05, NMR: dH1
=
3
The substrate specificity of the two a-fucosynthases was 5.37 ppm, JH1,H2 = 4.3 Hz): this unexpected compound, which
analyzed by using b-L-Fuc-N3 as donor in the presence of glyco- is present also in the reaction containing TmD224G, results as
side, monosaccharide, or disaccharide acceptors. Remarkably, trace impurity from the preparation of the donor.
SsD242S at donor:acceptor molar ratios between 1:2 and 1:3.4,
The relative amounts of the different transfucosylation prod-
catalyzed the synthesis of products from 16 different acceptors, ucts for each acceptor were obtained from the signal integration
including aryl-glycosides of hexoses, pentoses, N-acetyl- of the HPLC chromatogram while the amount of fucose trans-
glucosamine, disaccharides, and methylumbelliferyl-fuco- and ferred to an acceptor different from water was measured by
glucosides (Table S1). In all cases we observed also the forma- HPAEC-PAD (as described in Experimental Procedures),
tion of the autocondensation product 2. Interestingly, SsD242S yielding the global transfucosylation efficiency (Table 2). When
recognized as acceptors compounds containing either a/b-L/D inverted donor:acceptor molar ratios were used, such as 3:1 b-
anomeric bonds.
L-Fuc-N3: 4NP-b-D-Xyl, better total transfucosylation efficiency
Also TmD224G in the presence of 10 mM b-L-Fuc-N3 donor (76% versus 50%), but also higher amounts of a-L-Fuc-(1-3)-
and several acceptors (2- and 4NP-b-D-Xyl, 2NP-b-D-Fuc, and b-L-Fuc-N3 were found; therefore, an excess of donor was
4NP-b-D-Glc) used at 1:10 donor:acceptor molar ratios revealed considered detrimental for synthetic purposes.
several transfucosylation products, but we never observed the
autocondensation product (data not shown).
The regioselectivity of SsD242S depends on the acceptor. The
enzyme produced a-(1-3) and a-(1-4) linkages with 4NP-b-D-Xyl,
1100 Chemistry & Biology 16, 1097–1108, October 30, 2009 ª2009 Elsevier Ltd All rights reserved

Chemistry & Biology
a-L-Fucosynthases
Figure 2. Proposed Reaction Mechanisms of SsD242G/S Mutants in the Presence of 2C4NP-a-L-Fuc and Sodium Azide
The synthesis of a-L-Fuc-(1-3)-b-L-Fuc-N3 is explained assuming either that azide in b-L-Fuc-N3 donor works as leaving group leading to self-condensation (A) or
with the presence of a putative amino acid in the active site acting as novel nucleophile (B).
whereas with 4NP-b-D-Gal the regioselectivity was switched to 8 acts as acceptor. It is worth noting that compound 6, which
the formation of mainly a-(1-6) bond. Instead, with 4NP-b-D- is the most abundant product isolated in reaction II (78%) and
GlcNAc the a-(1-3) regioselectivity prevails with the formation contains an a-(1-6) linkage, did not lead to trisaccharides.
of a single product (10). This acceptor-dependent regioselectiv-
These results suggest that the +1 acceptor binding site of the
ity of exo-glycosynthases is not novel and has already been enzyme has high affinity for compound 8; by contrast, the auto-
described for the Streptomyces E384A b-glucosidase (Faijes condensation product 2 is a poor acceptor, as we never
et al., 2006).
observed trisaccharides of fucose. Interestingly, the transfuco-
Remarkably, with 4NP-b-D-Xyl and 4NP-b-D-Gal acceptors, sylation on disaccharide acceptors always led to branched prod-
the mutant catalyzed the formation of trisaccharides, namely ucts (compounds 5 and 9) and we never observed the transfer to
compounds 5 and 9. The former might be synthesized once the fucose at the nonreducing end of the acceptor.
compounds 3 or 4 compete with 1 in the acceptor subsite +1
The synthetic activity of TmD224G, tested on 4NP-b-D-Xyl
(for the GH active site nomenclature see Davies et al., 1997). acceptor (reaction IV), led to a-(1-4) and a-(1-3) linkage formation
Instead, the formation of the trisaccharide 9, which showed in about 1:1 molar ratio; however, this mutant is extremely effi-
a-(1-3) and a-(1-2), linkages, might occur only when compound cient in transfucosylations (91%).
Chemistry & Biology 16, 1097–1108, October 30, 2009 ª2009 Elsevier Ltd All rights reserved 1101

Chemistry & Biology
a-L-Fucosynthases
Figure 3. TLC Detection of Transfucosyla-
tion Activity of SsD242S and TmD224G
with b-L-fuc-azide Donor
(A) Lane 1: blank with no enzyme; lane 2: SsD242S
reaction; lane 3: fucose marker.
(B) Lane 1: fucose marker; lane 2: TmD224G (4 mg,
13 pmol) reaction, 1:5 donor:acceptor molar ratio;
lane 3 blank with no enzyme; lane 4: TmD224G
(1.3 mg, 4 pmol) reaction, 1:10; lane 5 blank; lane
6: as lane 4; lane 7: xylose marker; lanes 8 and 9
are the same as lanes 2 and 1 in (A), respectively.
The UV-visible product is indicated by an arrow-
head.
might facilitate the departure of the a-L-
Fuc-(1-2)-b-L-Fuc-N3 product from the
active site.
It is worth noting that in the crystalliza-
tion trials we never observed the forma-
Crystal Structure TmD224G in Complex with
an Autocondensation Product
tion of the complex of TmD224G with b-L-Fuc-N3. Presumably,
the mutation allowed the attack to the anomeric center of the
Soaking of TmD224G crystals in solutions containing up to 100 donor by an acceptor leading to the transfucosylation product.
mM b-L-Fuc-N3, followed by diffraction data analysis, did not Remarkably, in the crystal we observed only the transfucosyla-
reveal any electron density for the donor in the active site of tion product and not the hydrolytic product (L-fucose), which is
the mutant. We thus performed cocrystallization experiments efficiently bound by the wild-type enzyme (Sulzenbacher et al.,
by adding 75 mM b-L-Fuc-N3 to the crystallization buffer, ob- 2004). Possibly, this occurs because Gly224 in the mutant could
tained crystals after several weeks and collected diffraction not form the hydrogen bond to the 1-hydroxyl observed within
˚
data to 2.65 A resolution. Clear electron density could be the wild-type structure in complex with L-fucose (Sulzenbacher
observed in this case, but, to our surprise, it extended well et al., 2004).
beyond the donor binding site. Modeling of b-L-Fuc-N3 into
this electron density, in a position equivalent to b-L-fucose in DISCUSSION
the previously determined crystal structure of the Tma-fuc
product complex (Protein Data Bank code 1odu, Sulzenbacher Here we show that new mutants in the catalytic nucleophile of
et al., 2004), and subsequent refinement led to a very clean elec- two a-L-fucosidases from the hyperthermophiles S. solfataricus
tron density map around the pyranose ring, but left a substantial and T. maritima are efficient a-L-fucosynthases in the presence
amount of residual difference electron density around the azide of activated substrates and sodium azide. This ion usually
group. The only reasonable way to clean up the residual electron rescues the enzymatic activity of GH mutants in the nucleophile
density and make a good fit to the experimental data was to by acting as external nucleophile and producing glycosyl-azide,
model a disaccharide into the active site. This observation led which are stable and can be isolated easily and structurally char-
us to conclude that prolonged cocrystallization of TmD224G acterized (Ly and Withers, 1999; Zechel and Withers, 2001; Co-
with high concentrations of b-L-Fuc-N3 yielded an autoconden- bucci-Ponzano et al., 2003b; Shallom et al., 2005). In alternative
sation product, in contrast to what had been observed in solu- to azide, formate leads to glycosyl-formate intermediates,
tion. Unfortunately, the modest resolution of the diffraction which, being less stable than the glycosyl-azides, have been
data did not allow us to identify unambiguously the chemical exploited for the synthesis of oligosaccharides by retaining
nature of the autocondensation product. Models containing b-glycosynthases (Moracci et al., 1998; Perugino et al., 2004).
a-L-Fuc-(1-2/3/4)-b-L-Fuc-N3 regioisomers gave similar refine- Our initial attempts to use this approach on Ssa-fuc mutants
ment statistics, but the best fit could be obtained with a-L- failed, possibly because the b-L-fucosyl formate was not stable
Fuc-(1-2)-b-L-Fuc-N3. However, given the limited resolution, enough to act as intermediate. In contrast, the chemically
the great variety of possible conformational states (chair, boat, rescued activity of SsD242G/S mutants in the presence of
skew, etc.) of the pyranose ring at the reducing end could not sodium azide and 2C4NP-a-L-Fuc substrate led to the synthesis
be accounted for. In the model, which should be taken with of b-L-Fuc-N3 (1) and, more importantly, of the disaccharide
caution, the position of the fucose moiety at the nonreducing a-L-Fuc-(1-3)-b-L-Fuc-N3 (2), which could also be produced
end and its interactions with surrounding residues are essentially by SsD242S from b-L-Fuc-N3 donor. Therefore, SsD242G/S
the same as the ones observed in the structure of native Tma-fuc mutants can act as a-fucosynthase by following two different
in complex with L-fucose (Sulzenbacher et al., 2004). The fucosyl mechanisms depending on the reaction conditions: in the pres-
moiety at the reducing end makes no interactions with the ence of sodium azide and 2C4NP-a-L-Fuc substrate, the
enzyme, except for a stacking interaction between the azide mutants are able to perform in one pot both the reactions
group and Trp67 (Figure 4). Presumably, reduced interactions depicted in Figure 2A. Instead, when the mutants are incubated
1102 Chemistry & Biology 16, 1097–1108, October 30, 2009 ª2009 Elsevier Ltd All rights reserved

Chemistry & Biology
a-L-Fucosynthases
with suitable acceptors and b-L-Fuc-N3, they catalyzed transfu- presumably increases the mobility even more, possibly reorient-
cosylations (Figure 2A lower part). This is the reaction mecha- ing amino acids to make direct contact with the substrate and
nism followed by inverting glycosynthases (Figure S2A); the increasing the transfucosylation activity.
novelty here is that the donor is fucosyl azide rather than glyco-
side fluorides in conventional glycosynthases (for a review see homology model of Ssa-fuc, we observe striking differences
Williams and Withers, 2000). in the acceptor binding site. Tma-fuc residues Thr264 and
Glycosyl azides are substrates of transglycosylation reactions Arg254 are replaced by Ile287 and Val273, respectively: the
When we compare the crystal structure of Tma-fuc with a
´
catalyzed by retaining GHs (Fialova et al., 2005); mutants in the side chain of Ile287 makes only weak hydrophobic contact
catalytic nucleophile can use b-L-Fuc-N3 because of its stereo- with a nearby methionine, and consequently the loop region
chemistry and configuration. The higher stability of a-anomers carrying the Ssa-fuc acid/base Glu292 appears to be even less
(axial bonds) versus b-ones (equatorial bonds) is known as the anchored to the rest of the protein than in Tma-fuc. In subsite
‘‘anomeric effect’’ (Juaristi and Cuevas, 1992) explaining the +1 Tma-fuc Met225 is replaced in Ssa-fuc by Trp234, and this
much longer half-lives of a-glycosyl fluorides when compared latter residue might provide a better stacking platform for
with b-anomers (12 days versus 30 hr for a-D- and b-D-Glc F, pyranoside acceptors. Furthermore, a long insertion, missing in
respectively [Albert et al., 2000]). In addition, b-L-fucoside deriv- Ssa-fuc, is found in Tma-fuc after b strand b1, narrowing the
atives, with 1C4 conformation (compound 1 in Figure 2), being acceptor binding site. Residues found in this insertion are likely
6-deoxyhexopyranosides and showing axial substituents on to impose a steric and polar stringency on the nature of the
the C4, are more easily activable than hexopyranosides with acceptor sugar, not encountered in more open binding sites.
C4 equatorial substituents, respectively (Overend, 1972). There- These observations indicate that minor structural changes in
fore, notwithstanding the observation that glycosyl azides are the donor and acceptor binding sites dictate the glycosynthase
more stable than glycosyl fluorides, b-L-Fuc-N3 might be activity. Therefore, exploration of different strategies in order to
reactive enough to function as donor in the transfucosylation identify the best donor/acceptor substrates in a particular cata-
reactions catalyzed by the a-fucosynthases, requiring less lytic context is necessary.
acid/base assistance.
We have shown here that also the TmD224G mutant can act
as a-fucosynthase, confirming that the approach reported is
SIGNIFICANCE
of general application. The regiospecificity and the transfucosy-
lation efficiency differ for the two enzymes and depend on
the acceptor used. It is worth noting that the trisaccharide
fucosyloligosaccharides (compounds 5 and 9) and the disaccha-
ride a-L-Fuc-(1-3)-D-GlcNAc (compound 10), which is rather
common in glycoproteins and in the Lewisx, Lewisy, and Sialyl
Lewisx antigens, were easily produced by SsD242S with no
efforts to search for optimal reaction conditions. These results
suggest that a-L-fucosynthases might be further improved and
tailored to specific syntheses.
We report here the preparation of two novel retaining a-
fucosynthases, which follow the classical reaction pathway
proposed more than 10 years ago for b-glycosynthases
(Mackenzie et al., 1998), but utilize b-fucosyl azide as donor
substrate. Our findings might open new perspectives in the
use of azide derivatives for the production of novel a-glyco-
synthases. The only a-glucosynthase known so far utilizes
as donor the b-Glc-F (Okuyama et al., 2002), which has
a half-life of 30 hr in several aqueous buffer systems. More-
over, other glycoside fluoride derivatives are even less
stable: b-D-galactopyranoside- and b-D-mannopyranoside
fluorides have half-lives ranging between 6 and 8 hr and 11
and 17 hr, respectively (Albert et al., 2000). Therefore, fluori-
nated substrates in enzymatic a-galactosynthetic and
a-mannosynthetic reactions, often taking long incubations,
might be inappropriate substrates due to their enhanced
spontaneous hydrolysis at the operational conditions. In
these regards, azide derivatives might show the right
balance between stability and reactivity to work as suitable
donors in a-glycosynthetic reactions. Future work in the
development of novel glycosynthases should take into
account the relative stability of donors derivatized with
different chemical groups.
In the enzymes from S. solfataricus and T. maritima, only Ser
and Gly residues, respectively replacing the natural Asp nucleo-
phile, led to a a-fucosynthase. Ala residues did not work in both
enzymes, whereas, intriguingly, Ser, which best acted in the
former, did not function in Tma-fuc. In addition to this, the two
enzymes appear also to have different affinities for b-L-Fuc-N3
in the acceptor binding site. TmD224G forms an autocondensa-
tion product only at the high protein and substrate concentra-
tions found in the cocrystallization reaction, suggesting that
b-L-Fuc-N3 is a poor acceptor. Unfortunately, no electron
density could be observed beyond subsite +1, precluding further
analysis. Likewise, presuming disorder due to high mobility, no
electron density could be observed in the region succeeding
the acid/base Glu266, apparently important for transfucosylation
reactions. Indeed, a directed evolution study of Tma-fuc re-
ported that mutations Thr264Ala and Tyr267Phe enhance
considerably the transglycosylation activity of the enzyme
(Osanjo et al., 2007). Structural data show that the side chain
of Thr264 establishes a strong hydrogen bond with the side chain
of Arg254, whereas residues Ala265-His268 do not make
contact with any other residue of the protein, suggesting high
mobility in this region (Sulzenbacher et al., 2004). Abolishing
the anchoring of Thr264 to Arg254 by a Thr264Ala mutation
EXPERIMENTAL PROCEDURES
Chemicals
All commercially available substrates were purchased from Sigma-Aldrich.
The b-L-Fuc-N3 was chemically synthesized from L-Fuc in three steps and
67% yield according to a published procedure (Kunz et al., 1991). The Gene-
Tailor Site-directed Mutagenesis System was from Invitrogen; the synthetic
oligonucleotides (Table S2) were from PRIMM (Italy), and the His6 tagged
proteins were purified with the Protino Ni-TED 1000 protein purification system
(Macherey-Nagel, Germany).
Chemistry & Biology 16, 1097–1108, October 30, 2009 ª2009 Elsevier Ltd All rights reserved 1103

Chemistry & Biology
a-L-Fucosynthases
Table 2. Synthetic Products of SsD242S and TmD224G
Relative Ratios of
Transfucosylation
Products (%)
Reaction
I
Donor
Acceptor
Products
NDa
53
31
16
Total transfucosylation efficiency 50%
II
ND
78
10
1104 Chemistry & Biology 16, 1097–1108, October 30, 2009 ª2009 Elsevier Ltd All rights reserved

Chemistry & Biology
a-L-Fucosynthases
Table 2. Continued
Relative Ratios of
Transfucosylation
Products (%)
Reaction
Donor
Acceptor
Products
8
4
Total transfucosylation efficiency 26%
III
ND
100
Total transfucosylation efficiency 86%
IV
55
45
Total transfucosylation efficiency 91%
a ND, not determined.
Site-Directed Mutagenesis
2003b). Mutants SsD242A/S were prepared by site-directed mutagenesis
from the pGEX-frameFuc plasmid. Double mutants SsD242G/E58G-E292G
were prepared by polymerase chain reaction by using as template the vector
The plasmid pGEX-frameFuc expressing Ssa-fuc and the preparation of
SsD242G were described previously (Cobucci-Ponzano et al., 2003a,
Chemistry & Biology 16, 1097–1108, October 30, 2009 ª2009 Elsevier Ltd All rights reserved 1105

Chemistry & Biology
a-L-Fucosynthases
catalyzing the hydrolysis of 1 mmol substrate in 1 min at the conditions
described.
Steady-state kinetic parameters of Ssa-fuc and SsD242G/A/S mutants on
4NP-a-L-Fuc and 2C4NP-a-L-Fuc were described previously (Cobucci-
Ponzano et al., 2003b, 2008). Steady-state kinetic parameters of wild-type
Tma-fuc and TmD224G were measured in buffer B at 60ꢁC on 4NP-a-L-Fuc
and 2C4NP-a-L-Fuc used at concentrations ranges of 0.005–0.8 mM, and,
where indicated, in the presence of 1 M sodium azide. The amount of the
enzyme used in the assays was 5 mg (16 pmol).
In all the enzymatic assays, spontaneous hydrolysis of the substrate was
subtracted by using appropriate blank mixtures without enzyme. All kinetic
data were calculated as the average of at least two experiments and were
plotted and refined with the program GraFit (Leatherbarrow, 1992).
Glycosynthetic Trials
To purify and characterize the glycosynthetic products, SsD242G (50 mg,
100 pmol) was incubated for 16 hr at 65ꢁC in 0.8 ml buffer A in the presence
of 0.1 M sodium azide, and 20 mM 2C4NP-a-L-Fuc. The transfucosylation
products of SsD242S from b-L-Fuc-N3 donor to different acceptors were
prepared by incubating 94 mg (188 pmol) enzyme for 16 hr at 65ꢁC in 0.8 ml
buffer A, by using different donor:acceptor molar ratios. The transfucosylation
products of TmD224G were prepared by incubating 38 mg (122 pmol) enzyme
for 16 hr at 70ꢁC in 0.2 ml buffer A by using 10 mM b-L-Fuc-N3 donor and
100 mM 4NP-b-D-Xyl (1:10 molar ratio). Blank mixtures without enzyme
were also prepared. 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% a-naphthol in 10% sulfuric acid in ethanol
followed by charring.
Figure 4. Close-up View of the Active Site of TmD224G
a-L-Fuc-(1-2)-b-L-Fuc-N3 is shown in stick representation with carbon,
oxygen, and nitrogen atoms colored in yellow, red, and blue, respectively.
Residues interacting with the disaccharide are shown as sticks under a trans-
parent surface, with carbon, oxygen, and nitrogen atoms colored in gray, red,
and blue, respectively. A maximum-likelihood/sA-weighted Fo-Fc electron
density calculated prior incorporation of the disaccharide into the model and
contoured at 2.5 s is shown in green.
The reaction mixtures were separately frozen dried and the purification of
each sample was obtained by reverse-phase chromatography (Polar-RP
80A, Phenomenex, 4 m, 250 3 10 mm) on an Agilent HPLC instrument 1100
series, using 3:2 water/methanol as eluent. The eluted products were first
analyzed by positive ions reflection MALDI-TOF mass spectrometry. For char-
acterization of the products, see Supplemental Experimental Procedures.
Analysis of transfucosylation efficiency of Ssa-fuc and Tma-fuc mutants,
was performed by use of a high-performance anion-exchange chromatog-
raphy with pulsed amperometric detection (HPAEC-PAD) equipped with a
PA1 column (Dionex, USA). The reaction mixtures and the blank mixtures
described above werediluted(10-to 100-fold) withH2O and 0.5 nmol arabinose
as internal standard, loaded onto the PA1 column and eluted with 16 mM
NaOH. The moles of fucose were determined by integration of the peaks within
the chromatogram, based on fucose and arabinose standard curves. The
amount of fucose transferred by the enzyme to water was calculated by sub-
tracting the amount of free fucose measured in the blank mixtures from that
identified in the reaction mixtures.
expressing SsD242G and the mutagenic oligonucleotides reported previously
(Cobucci-Ponzano et al., 2005).
The plasmid expressing Tma-fuc was described previously (Sulzenbacher
et al., 2004) and it was used as template for the preparation by site-directed
mutagenesis of TmD224G and TmD224S. The genes containing the desired
mutations were identified by direct sequencing and completely resequenced.
Expression and Purification of Mutant a-L-Fucosidases
The mutant a-L-fucosidases from S. solfataricus were expressed as previously
described and purified by a slight modification of the final heating steps
at 65ꢁC and 70ꢁC (Cobucci-Ponzano et al., 2003b). The mutants of the
T. maritima a-L-fucosidases were expressed as previously reported (Tarling
et al., 2003). The purification procedures were performed with affinity matrixes
dedicated only to the purification of the specific mutant to exclude contamina-
tion by the wild-type enzyme from external sources. The enzymes, > 95% pure
by SDS-PAGE, stored at 4ꢁC in sodium phosphate buffer 20 mM (pH 7.0), NaCl
150 mM (phosphate-buffered saline) were stable for several months at 4ꢁC.
Protein concentration was determined with the method of Bradford (Bradford,
1976), by using bovine serum albumin as standard.
To measure the total amount of fucose enzymatically transferred, 1/10 of the
reaction mixtures were incubated for 90 min at 65ꢁC in the presence of 1.2 mg
(2.4 pmol) Ssa-fuc wild-type. Successively, the solution was treated as
described above and run by HPAEC-PAD to measure the total amount of
fucose. The efficiency of the transfucosylation reaction was calculated as: total
amount of fucose transferred ꢀ moles of fucose transferred to water / total
amount of fucose transferred 3 100.
Crystallographic Analysis
Crystals of TmD224G in complex with b-L-Fuc-N3 were obtained as reported
for the native enzyme (Sulzenbacher et al., 2004), with the addition of 14 mg/
ml b-L-Fuc-N3 to the crystallization solution. Crystals belong to space group
H32 and contain two molecules of TmD224G in the asymmetric unit. Diffrac-
Enzymatic Characterization
The activity of the wild-type and mutant Ssa-fuc at standard conditions was
measured at 65ꢁC in 50 mM sodium phosphate buffer (pH 6.5, buffer A),
with 4NP-a-L-Fuc (1 mM) and 2C4NP-a-L-Fuc (3 mM) by using up to 20 mg
(40 pmol) enzyme. The activity of TmD224G was measured on the same
substrates and concentrations at 60ꢁC in 50 mM sodium citrate/phosphate
buffer (pH 6.0, buffer B).
˚
tion data extending to 2.65 A were collected from a flash-frozen crystal at the
ESRF beam line ID14-2. Data were indexed and integrated with MOSFLM
(Leslie, 1992) and all further computing was carried out with the CCP4
program suite (CCPN, 1994), unless otherwise stated. Data collection statis-
tics are summarized in Table 3. The structure of TmD224G in complex with
b-L-Fuc-N3 was solved by molecular replacement with the program PHASER
(McCoy et al., 2007), using the crystal structure of native Tma-fuc (Protein
Data Bank code 1hl8) as a search model, and the resulting model was
refined with REFMAC (Murshudov et al., 1997) using the maximum like-
lihood approach and incorporating bulk solvent corrections and anisotropic
The chemically rescued activities of SsD242S and TmD242G on the same
substrates, in the presence of sodium formate or sodium azide, were
measured at the indicated conditions.
The molar extinction coefficients used are reported in Table S3. For all these
enzymes, one unit of enzyme activity was defined as the amount of enzyme
1106 Chemistry & Biology 16, 1097–1108, October 30, 2009 ª2009 Elsevier Ltd All rights reserved

Chemistry & Biology
a-L-Fucosynthases
Received: July 17, 2009
Table 3. Crystallographic Data Collection and Refinement
Statistics
Revised: September 8, 2009
Accepted: September 14, 2009
Published: October 30, 2009
Data
Wavelength
0.933
˚
a = b, c (A)
Resolution rangea
a,b
Rmerge
180.25, 169.73
35 – 2.65 (2.79 – 2.65)
0.080 (0.455)
207721
REFERENCES
Aharoni, A., Thieme, K., Chiu, C.P., Buchini, S., Lairson, L.L., Chen, H., Stry-
nadka, N.C., Wakarchuk, W.W., and Withers, S.G. (2006). High-throughput
screening methodology for the directed evolution of glycosyltransferases.
Nat. Methods 3, 609–614.
No. of observations
No. of unique reflections
Completeness (%)a
Redundancya
30789
Albert, M., Repetschnigg, W., Ortner, J., Gomes, J., Paul, B.J., Illaszewicz, C.,
Weber, H., Steiner, W., and Dax, K. (2000). Simultaneous detection of different
glycosidase activities by 19F NMR spectroscopy. Carbohydr. Res. 327,
395–400.
99.9 (100.0)
6.7 (6.5)
<I/sI>a
16.5 (3.8)
B from Wilson statistics
Refinement
65.3
Ben-David, A., Shoham, G., and Shoham, Y. (2008). A universal screening
assay for glycosynthases: directed evolution of glycosynthase XynB2(E335G)
suggests a general path to enhance activity. Chem. Biol. 15, 546–551.
˚
Resolution (A)
35-2.65
Berman, H.M., Westbrook, J., Feng, Z., Gilliland, G., Bhat, T.N., Weissig, H.,
Shindyalov, I.N., and Bourne, P.E. (2000). The Protein Data Bank. Nucleic
Acids Res. 28, 235–242.
No. of protein atomsc
No. of water molecules/ligand atomsc
Rcrystd / Rfree (%)
7168
103 / 46
20.14 / 24.06
0.006 (0.020)
0.949 (1.95)
52.15 / 53.07
24.37 / 60.26
0.77 / 1.14
Bradford, M.M. (1976). A rapid and sensitive method for the quantitation of
microgram quantities of protein utilizing the principle of protein-dye binding.
Anal. Biochem. 72, 248–254.
˚
Rmsd 1-2 bond distances (A)
Rmsd 1-3 bond angles (ꢁ)
2
Average main/side chain B (A )
2
Average B solvent / ligand (A )
Cantarel, B.L., Coutinho, P.M., Rancurel, C., Bernard, T., Lombard, V., and
Henrissat, B. (2009). The Carbohydrate-Active EnZymes database (CAZy):
an expert resource for Glycogenomics. Nucleic Acids Res. 10.1093/nar/
gkn663.
˚
˚
Main chain / side chain D B, bonded
2
atoms (A )
˚
CCP4 (Collaborative Computational Project, Number 4). (1994). The CCP4
suite: Programs for protein crystallography. Acta Crystallogr. D Biol. Crystal-
logr. 50, 760–763.
a Values in parentheses are for the highest resolution shell.
b Rmerge = Shkl Si j Ihkli - <Ihkli>j/ Shkl Si <Ihkli>.
c Per asymmetric unit, corresponding to 2 molecules of TmD224G. Target
mean values and standard deviations are given in parentheses.
d Rcryst = SkFoj - jFck/SjFoj.
Cobucci-Ponzano, B., Trincone, A., Giordano, A., Rossi, M., and Moracci, M.
(2003a). Identification of an archaeal alpha-L-fucosidase encoded by an
interrupted gene. Production of a functional enzyme by mutations mimicking
programmed –1 frameshifting. J. Biol. Chem. 278, 14622–14631.
Cobucci-Ponzano, B., Trincone, A., Giordano, A., Rossi, M., and Moracci, M.
(2003b). Identification of the catalytic nucleophile of the family 29 alpha-L-
fucosidase from Sulfolobus solfataricus via chemical rescue of an inactive
mutant. Biochemistry 42, 9525–9531.
Fobs versus Fcalc scaling. A random 5% (1536) of reflections were set aside
for cross-validation purposes. Manual adjustments of the model were carried
out with Coot (Emsley and Cowtan, 2004). Refinement and structure quality
statistics are listed in Table 3. Coordinates have been deposited in the
Protein Data Bank (Berman et al., 2000) with accession reference number
2wsp. Figure 4 was generated using the program PYMOL (http://pymol.
sourceforge.net/) (DeLano, 2002). A homology model of Ssa-fuc was gener-
ated with the Phyre Server (http://www.sbg.bio.ic.ac.uk/phyre/) (Kelley and
Sternberg, 2009).
Cobucci-Ponzano, B., Mazzone, M., Rossi, M., and Moracci, M. (2005).
Probing the catalytically essential residues of the alpha-L-fucosidase from
the hyperthermophilic archaeon Sulfolobus solfataricus. Biochemistry 44,
6331–6342.
Cobucci-Ponzano, B., Conte, F., Mazzone, M., Bedini, E., Corsaro, M.M.,
Rossi, M., and Moracci, M. (2008). Design of new reaction conditions for
characterization of a mutant thermophilic a-L-fucosidase. Biocat. Biotrans.
26, 18–24.
ACCESSION NUMBERS
Davies, G.J., Wilson, K.S., and Henrissat, B. (1997). Nomenclature for sugar-
Coordinates have been deposited in the Research Collaboratory for Structural
Bioinformatics Protein Data Bank with the accession code 1odu.
binding subsites in glycosyl hydrolases. Biochem. J. 321, 557–559.
DeLano, W.L. (2002). The PyMOL molecular graphics system (San Carlos, CA:
DeLano Scientific).
Ducros, V.M., Tarling, C.A., Zechel, D.L., Brzozowski, A.M., Frandsen, T.P.,
von Ossowski, I., Schulein, M., Withers, S.G., and Davies, G.J. (2003).
Anatomy of glycosynthesis: structure and kinetics of the Humicola insolens
Cel7B E197A and E197S glycosynthase mutants. Chem. Biol. 10, 619–628.
SUPPLEMENTAL DATA
Supplemental Data include four figures, three tables, and Supplemental
Experimental Procedures, and can be found with the article online at http://
www.cell.com/chemistry-biology/supplemental/S1074-5521(09)00322-6.
Emsley, P., and Cowtan, K. (2004). Coot: model-building tools for molecular
graphics. Acta Crystallogr. D Biol. Crystallogr. 60, 2126–2132.
´
Faijes, M., Saura-Valls, M., Perez, X., Conti, M., and Planas, A. (2006).
ACKNOWLEDGMENTS
Acceptor-dependent regioselectivity of glycosynthase reactions by Strepto-
myces E383A beta-glucosidase. Carbohydr. Res. 341, 2055–2065.
We thank Francesca De Micco for kinetic studies of Tma-fuc. This work was
supported by the project MoMa n. 1/014/06/0 of the Agenzia Spaziale Italiana.
The IBP-CNR belongs to the Centro Regionale di Competenza in Applicazioni
Tecnologico-Industriali di Biomolecole e Biosistemi.
Farkas, E., Thiem, J., and Ajisaka, K. (2000). Enzymatic synthesis of fucose-
containing disaccharides employing the partially purified alpha-L-fucosidase
from Penicillium multicolor. Carbohydr. Res. 328, 293–299.
Chemistry & Biology 16, 1097–1108, October 30, 2009 ª2009 Elsevier Ltd All rights reserved 1107

Chemistry & Biology
a-L-Fucosynthases
´
Fialova, P., Carmona, A.T., Robina, I., Ettrich, R., Sedmera, P., Prikrylova, V.,
´
Murshudov, G.N., Vagin, A.A., and Dodson, E.J. (1997). Refinement of macro-
molecular structures by the maximum-likelihood method. Acta Crystallogr.
D Biol. Crystallogr. 53, 240–255.
´
Petraskova-Husakova, L., and Kren, V. (2005). Glycosyl azide—a novel
´
´
substrate for enzymatic transglycosylations. Tetrahedron Lett. 46, 8715–8718.
Okuyama, M., Mori, H., Watanabe, K., Kimura, A., and Chiba, S. (2002).
a-glucosidase mutant catalyzes ‘‘a-glycosynthase’’-type reaction. Biosci.
Biotechnol. Biochem. 66, 928–933.
Hancock, S.M., Vaughan, M.D., and Withers, S.G. (2006). Engineering of
glycosidases and glycosyltransferases. Curr. Opin. Chem. Biol. 10, 509–519.
Honda, Y., and Kitaoka, M. (2006). The first glycosynthase derived from an
Osanjo, G., Dion, M., Drone, J., Solleux, C., Tran, V., Rabiller, C., and Tellier, C.
(2007). Directed evolution of the alpha-L-fucosidase from Thermotoga mari-
tima into an alpha-L-transfucosidase. Biochemistry 46, 1022–1033.
inverting glycoside hydrolase. J. Biol. Chem. 281, 1426–1431.
Honda, Y., Fushinobu, S., Hidaka, M., Wakagi, T., Shoun, H., Taniguchi, H.,
and Kitaoka, M. (2008). Alternative strategy for converting an inverting glyco-
side hydrolase into a glycosynthase. Glycobiology 18, 325–330.
Overend, W.G. (1972). In The Carbohydrates, 2nd Edition, W. Pigman and
D. Horton, eds. (New York and London: Academic Press), pp. 316–332.
Juaristi, E., and Cuevas, G. (1992). Recent studies of the anomeric effect.
Perugino, G., Trincone, A., Rossi, M., and Moracci, M. (2004). Oligosaccharide
Tetrahedron 48, 5019–5087.
synthesis by glycosynthases. Trends Biotechnol. 22, 31–37.
Kelley, L.A., and Sternberg, M.J.E. (2009). Protein structure prediction on the
Perugino, G., Cobucci-Ponzano, B., Rossi, M., and Moracci, M. (2005). Recent
advances in the oligosaccharide synthesis promoted by catalytically engi-
neered glycosidases. Adv. Synth. Catal. 347, 941–950.
web: a case study using the Phyre server. Nat. Protoc. 4, 363–371.
Kim, Y.W., Lee, S.S., Warren, R.A., and Withers, S.G. (2004). Directed evolu-
tion of a glycosynthase from Agrobacterium sp. increases its catalytic activity
dramatically and expands its substrate repertoire. J. Biol. Chem. 279,
42787–42793.
Seeberger, P.H. (2008). Automated oligosaccharide synthesis. Chem. Soc.
Rev. 37, 19–28.
Shallom, D., Leon, M., Bravman, T., Ben-David, A., Zaide, G., Belakhov, V., Sho-
ham, G., Schomburg, D., Baasov, T., and Shoham, Y. (2005). Biochemical char-
acterization and identification of the catalytic residues of a family 43 beta-D-
xylosidase from Geobacillus stearothermophilus T-6. Biochemistry 44, 387–397.
Kunz, H., Pfrengle, W., Ru¨ ck, K., and Sager, W. (1991). Stereoselective
synthesis of L-amino acids via Strecker and Ugi reactions on carbohydrate
templates. Synthesis 1991, 1039–1042.
Leatherbarrow, R.J. (1992). GraFit version 3.0 (Staines, U.K.: Erithacus Soft-
ware Ltd.).
Sulzenbacher, G., Bignon, C., Nishimura, T., Tarling, C.A., Withers, S.G., Hen-
rissat, B., and Bourne, Y. (2004). Crystal structure of Thermotoga maritima
alpha-L-fucosidase. Insights into the catalytic mechanism and the molecular
basis for fucosidosis. J. Biol. Chem. 279, 13119–13128.
Leslie, A.G.W. (1992). Recent changes to the MOSFLM package for process-
ing film and image plate data. Joint CCP4/ESF-EACBM Newsletter 26, 27–33.
Tarling, C.A., He, S., Sulzenbacher, G., Bignon, C., Bourne, Y., Henrissat, B.,
and Withers, S.G. (2003). Identification of the catalytic nucleophile of the family
29 alpha-L-fucosidase from Thermotoga maritima through trapping of a cova-
lent glycosyl-enzyme intermediate and mutagenesis. J. Biol. Chem. 278,
47394–47399.
Lin, H., Tao, H., and Cornish, V.W. (2004). Directed evolution of a glycosyn-
thase via chemical complementation. J. Am. Chem. Soc. 126, 15051–15059.
Ly, H.D., and Withers, S.G. (1999). Mutagenesis of glycosidases. Annu. Rev.
Biochem. 68, 487–522.
Ma, B., Simala-Grant, J.L., and Taylor, D.E. (2006). Fucosylation in prokaryotes
Umekawa, M., Huang, W., Li, B., Fujita, K., Ashida, H., Wang, L.X., and Yama-
moto, K. (2008). Mutants of Mucor hiemalis endo-beta-N-acetylglucosamini-
dase show enhanced transglycosylation and glycosynthase-like activities.
J. Biol. Chem. 283, 4469–4479.
and eukaryotes. Glycobiology 16, 158R–184R.
Mackenzie, L.F., Wang, Q., Warren, R.A.J., and Withers, S.G. (1998). Glyco-
synthases: mutant glycosidases for oligosaccharide synthesis. J. Am. Chem.
Soc. 120, 5583–5584.
Vanhooren, P.T., and Vandamme, E.J. (1999). L-Fucose: occurrence, physio-
logical role, chemical, enzymatic and microbial synthesis. J. Chem. Technol.
Biotechnol. 74, 479–497.
Malet, C., and Planas, A. (1998). From beta-glucanase to beta-glucansyn-
thase: glycosyl transfer to alpha-glycosyl fluorides catalyzed by a mutant
endoglucanase lacking its catalytic nucleophile. FEBS Lett. 440, 208–212.
Varki, A. (1993). Biological roles of oligosaccharides: all of the theories are
McCarter, J.D., and Withers, S.G. (1994). Mechanisms of enzymatic glycoside
correct. Glycobiology 3, 97–130.
hydrolysis. Curr. Opin. Struct. Biol. 4, 885–892.
Wada, J., Honda, Y., Nagae, M., Kato, R., Wakatsuki, S., Katayama, T.,
Taniguchi, H., Kumagai, H., Kitaoka, M., and Yamamoto, K. (2008). 1,2-
alpha-l-Fucosynthase: A glycosynthase derived from an inverting alpha-
glycosidase with an unusual reaction mechanism. FEBS Lett. 582, 3739–3743.
McCoy, A.J., Grosse-Kunstleve, R.W., Adams, P.D., Winn, M.D., Storoni, L.C.,
and Read, R.J. (2007). Phaser Crystallographic Software. J. Appl. Crystallogr.
40, 658–674.
Williams, S.J., and Withers, S.G. (2000). Glycosyl fluorides in enzymatic reac-
Moracci, M., Trincone, A., Perugino, G., Ciaramella, M., and Rossi, M. (1998).
Restoration of the activity of active-site mutants of the hyperthermophilic beta-
glycosidase from Sulfolobus solfataricus: dependence of the mechanism on
the action of external nucleophiles. Biochemistry 37, 17262–17270.
tions. Carbohydr. Res. 327, 27–46.
Williams, G.J., Zhang, C., and Thorson, J.S. (2007). Expanding the promiscuity
of a natural-product glycosyltransferase by directed evolution. Nat. Chem.
Biol. 3, 657–662.
Murata, T., Morimoto, S., Zeng, X., Watanabe, S., and Usui, T. (1999). Enzy-
matic synthesis of alpha-L-fucosyl-N-acetyllactosamines and 30-O-alpha-L-
fucosyllactose utilizing alpha-L-fucosidases. Carbohydr. Res. 320, 192–199.
Zechel, D.L., and Withers, S.G. (2001). Dissection of nucleophilic and acid-
base catalysis in glycosidases. Curr. Opin. Chem. Biol. 5, 643–649.
1108 Chemistry & Biology 16, 1097–1108, October 30, 2009 ª2009 Elsevier Ltd All rights reserved