Family 4 glycosidases carry out efficient hydrolysis of thioglycosides by an α,β-elimination mechanism

Article abstract of DOI:10.1002/anie.200601421

Thioglycosides are widely used as competitive inhibitors of glycosidases. The first glycosidase capable of hydrolyzing unactivated thioglycosides - BglT, a 6-phospho-β-glucosidase from glucosidase family 4 - has been found. The efficient cleavage of the t

Full text of DOI:10.1002/anie.200601421

Angewandte
Chemie
Thioglycoside Hydrolysis
DOI: 10.1002/anie.200601421
Family 4 Glycosidases Carry Out Efficient
Hydrolysis of Thioglycosides by an a,b-
Elimination Mechanism**
Vivian L. Y. Yip and Stephen G. Withers*
Considerable efforts have been extended towards the devel-
opment of glycosidase inhibitors, both as stable substrate
analogues for structural and mechanistic studies and for
potential therapeutic and industrial applications.[1,2] Amongst
those developed, thioglycosides, in which the glycosidic
oxygen has been replaced by a sulfur atom, have proved to
be stable analogues of the ground-state substrate and have
been employed in a number of insightful structural stud-
ies.[3,4][ ] Glycosidases are known to effect hydrolysis by acid/
*
base-catalyzed mechanisms involving oxocarbenium-ion-like
transition states.[5–7] Consequently, the resistance of the
thioglycosidic bond to cleavage has been ascribed to the
lower proton affinity of sulfur over that of oxygen, resulting in
inefficient general acid catalysis to the departing agly-
cone.[8–10]
[*] V. L. Y. Yip, Prof. S. G. Withers
Department of Chemistry
University of BritishColumbia
2036 Main Mall, Vancouver, BC V6T1Z1 (Canada)
Fax: (+1)604-822-8869
E-mail: withers@chem.ubc.ca
[**] We thank Dr. Gideon Davies for providing the BglTplasmid, Dr. John
Thompson for providing the kinase BglK, Dr. J. Müllegger for
providing the thioglycoligase Abg E170Q, and Dr. Hongming Chen
for providing p-nitrophenyl 4-deoxy-4-thio-d-glucoside. We also
thank the Natural Sciences and Engineering Research Council of
Canada for financial support. V.L.Y.Y. is funded by graduate
fellowships from the Natural Sciences and Engineering Research
Council of Canada and the Michael Smith Foundation for Health
Research.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
[*] The one exception is a specialized group of S-glycosidases of plant
origin, called the myrosinases (E.C. 3.2.3.1), which specifically
hydrolyze glucosinolate substrates, anionic 1-thio-b-glucosides. By
sequence aligment, the myrosinases are associated with the family 1
glycosidases, which catalyze the hydrolysis of b-O-glycosides with
retention of the substrate anomeric configuration. While the majority
of family 1 glycosidases contain a conserved glutamate as the
catalytic nucleophile, in the myrosinases the glutamic acid that
serves as the acid/base catalyst is replaced by a glutamine residue. It
has been proposed that the myrosinases are able to catalyze the
cleavage of glucosinolate, because the substrates contain inherently
good leaving groups and thus do not require general acid assistance
for leaving-group departure.[25] A bound ascorbate anion appears to
function as the general base catalyst for hydrolysis of the glycosyl–
enzyme intermediate. A similar explanation is given for the GH84
human O-GlcNAcase, which is proposed to catalyze the hydrolysis of
activated thioglycosides via a very dissociative transition state and
without general acid catalysis to protonate the thiolate leaving
group.[21]
Angew. Chem. Int. Ed. 2006, 45, 6179 –6182
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
6179

Communications
The enzymes that have evolved to cleave carbon–sulfur
can efficiently hydrolyze thioglycosidic bonds. Not only is this
of fundamental mechanistic interest, but also it is highly
relevant to the design of specific glycosidase inhibitors as
potential therapeutics. Preliminary results[19] had indicated
that thioglycosides might, in general, function as substrates
linkages generally appear to use an anionic elimination
mechanism. Excellent examples include cysteine Cb-Sg
lyases,[11] b-cystathionase,[12] S-(b-aminoethyl)-cysteine,[13] S-
alkylcysteine lyase,[14,15] S-adenosyl homocysteine hydrolase
(AdoHCyase),[16] and the more recently uncovered ribosyl
homocysteinase.[17] Most significantly, AdoHCyase[16] utilizes
an NAD+ cofactor to oxidize the ribose C3hydroxy group of
S-adenosyl homocysteine, thereby lowering the pKa of the C4
proton and facilitating deprotonation and hence a,b-elimi-
nation of the thiol. The Michael acceptor then undergoes 1,4
nucleophilic attack by a water molecule, followed by reduc-
tion of the C3ketone by the “on-board” NADH, yielding
adenosine and homocysteine and returning the enzyme to its
initial catalytic state.
since
p-nitrophenyl
6-phospho-b-d-thioglucoside
(S-
pNPG6P) was shown to be hydrolyzed with kinetic param-
eters that are very similar to those of its oxygen-containing
counterpart, p-nitrophenyl 6-phospho-b-d-glucoside (O-
pNPG6P).[19] However, the use of an activated aryl leaving
group could be misleading in this case as previous studies with
a “normal” family 1 b-glucosidase, Abg,[20] and a GH84
human O-GlcNAcase[21] have shown that activated aryl
thioglycosides could indeed be cleaved reasonably efficiently.
To properly test whether family 4 enzymes can cleave
thioglycosides, it was necessary to study the hydrolysis of
non-activated thiodisaccharide substrates. Accordingly, we
set out to synthesize and test thioglycoside analogues of the
natural substrate for the Thermotoga maritima BglT, a 6-
phospho-b-glucosidase from GH4 for which a three-dimen-
sional structure and a number of mechanistic studies are
available. To simplify assays, as well as synthetic routes, we
elected to synthesize the p-nitrophenyl glycoside of thiocel-
lobiose, as well as its oxygen-linked counterpart.
We recently unveiled a completely new mechanism of
enzymatic glycoside hydrolysis analogous to that of AdoH-
Cyase and also involving anionic transition states. The
glycosidases that employ this mechanism are found within
glycoside hydrolase family 4 (GH4), which also use an “on-
board” NAD+. In this case, the cofactor is used to transiently
[18]
oxidize the C3hydroxy group, as shown in Scheme 1.
Key
features of the mechanism include oxidation of the C3
hydroxy group, formation of an enediolate intermediate
stabilized by a bound Mn2+ cofactor, addition of water to
the Michael acceptor so formed, and finally reduction of the
ketone.
The similarities of the elimination mechanisms of AdoH-
Cyase and glycosidases from family 4 raised the question as to
whether GH4 enzymes, in contrast to all other glycosidases,
p-Nitrophenyl 4-deoxy-4-thio-6’-phospho-b-d-cellobio-
side (S-pNPC6’P) was synthesized according to the chemo-
enzymatic route shown in Scheme 2 starting with p-nitro-
phenyl 4-deoxy-4-thio-b-d-glucoside. The key step involved
an enzymatic coupling using the thioglycoligase technology
recently developed in our group to produce the thiodisac-
Scheme 1. Proposed mechanism of BglT, a 6-phospho-b-glucosidase from GH4.
6180 www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 6179 –6182

Angewandte
Chemie
dehydrogenase. A second, cou-
pled assay system was also
employed in which the p-nitro-
phenyl b-d-glucoside released
from O-pNPC6’P was hydrolyzed
by an added b-glucosidase. How-
ever, since
p-nitrophenyl 4-deoxy-4-thio-b-d-
glucoside is not a substrate for
Abg, this coupled assay could not
be used for S-pNPC6’P.
The kinetic parameters for the
hydrolysis of the disaccharide sub-
strates are shown in Table 1 along
with Ki values for each disacchar-
ide as an inhibitor of O-pNPG6P
hydrolysis. The kinetic parame-
ters for the disaccharide sub-
strates are similar; the kcat and
KM values for S-pNPC6’P are
larger than those for O-pNPC6’P,
resulting in similar kcat/KM values.
Meanwhile, the kcat and KM values
Scheme 2. Synthesis of S-pNPC6’P and S-C6’P.
charide p-nitrophenyl 4-deoxy-4-thio-b-d-cellobioside (S-
pNPC).[22] The free disaccharide, 4-deoxy-4-thio-d-cellobiose
(S-C; Scheme 2), was prepared from S-pNPC by removal of
the p-nitrophenyl group at C1 using the cellulase Onozuka R-
10 from Trichoderma viride. The disaccharide products S-
pNPC and S-C were then selectively phosphorylated at O6’
using the kinase BglK from Klebsiella pneumoniae that is
associated with the phosphoenolpyruvate-dependent sugar:
phosphotransferase system (PEP:PTS) used by many bacte-
rial species for the simultaneous phosphorylation and trans-
location of carbohydrates into the cell.[23] The oxygen
analogues, p-nitrophenyl 6’-phospho-b-d-cellobioside (O-
pNPC6’P) and d-cellobiose 6’-phosphate (O-C6’P), were
synthesized analogously from the p-nitrophenyl b-d-cellobio-
side and d-cellobiose, respectively, using the same kinase
(Scheme 3).
Table 1: Summary of kinetic parameters for the hydrolysis of O- and S-
glucosides by BglT.
Substrate
k
cat [sÀ1
]
KM [mm]
kcat/KM
Ki [mm]
[sÀ1 mmÀ1
]
O-pNPG6P[a]
S-pNPG6P[a]
O-pNPC6’P
S-pNPC6’P
O-C6’P[b]
1.9
1.6
0.017
0.072
0.61
0.53
41
41
1.0
8.0
69
37
46
39
17
9.0
8.8
14
–
–
1.0
5.0
70
33
S-C6’P
[a] Data obtained from Ref. [19]. [b] Data obtained from Ref. [24].
for O-C6’P are both slightly larger than those for S-C6’P,
resulting in similar kcat/KM values. The similarities of the KM
and Ki values in each case confirms that the reaction is indeed
occurring through the same active site. Furthermore, product
analyses by NMR spectroscopy and mass spectrometry
showed that the expected products (glucose 6-phosphate
and 4-deoxy-4-thioglucose) are indeed formed when S-C6’P is
reacted with BglT. In addition, deuterium
incorporation from the solvent into the
glucose 6-phosphate product strongly sug-
gests that BglTutilizes the same mechanism
in the hydrolysis of thiodisaccharides as that
of O-glycosidic linkages. Overall, there is
no significant difference between the rates
of hydrolysis of O- and S-glycosidic linkages
as catalyzed by BglT. Hence, we have
shown that BglT is the first glycosidase
capable of efficiently cleaving unactivated
thioglycosides.
Kinetic parameters for all substrates were determined
using a coupled assay system, in which the formation of the
glucose 6-phosphate product was coupled to the reduction of
NADP through inclusion of the enzyme glucose 6-phosphate
These findings clearly indicate a funda-
mental difference in the BglT mechanism
compared to that of most glycosidases,
Scheme 3. Disaccharide substrates for BglT.
Angew. Chem. Int. Ed. 2006, 45, 6179 –6182
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
6181

Communications
[12] C. R. Gentry-Weeks, J. Spokes, J. Thompson, J. Biol. Chem.
1995, 270, 7695.
[13] I. Rossol, A. Puhler, J. Bacteriol. 1992, 174, 2968.
[14] H. Kamitani, N. Esaki, H. Tanaka, K. Soda, Agric. Biol. Chem.
1990, 54, 2069.
[15] C.-H. Tai, P. F. Cook, Acc. Chem. Res. 2001, 34, 49.
[16] J. L. Palmer, R. H. Abeles, J. Biol. Chem. 1979, 254, 1217.
[17] D. Pei, J. Zhu, Curr. Opin. Chem. Biol. 2004, 8, 492.
[18] V. L. Y. Yip, A. Varrot, G. J. Davies, S. S. Rajan, X. J. Yang, J.
Thompson, W. F. Anderson, S. G. Withers, J. Am. Chem. Soc.
2004, 126, 8354.
[19] A. Varrot, V. L. Y. Yip, Y. Li, S. S. Rajan, X. Yang, W. F.
Anderson, J. Thompson, S. G. Withers, G. J. Davies, J. Mol. Biol.
2005, 346, 423.
[20] A. G. Day, Purification and Preliminary Characterization of b-
Glucosidase from Alcaligenes faecalis (ATCC 21400), MSc
Thesis, 1985.
[21] M. S. Macauley, K. A. Stubbs, D. J. Vocadlo, J. Am. Chem. Soc.
2005, 127, 17202.
[22] J. Mullegger, M. Jahn, H. Chen, R. A. J. Warren, S. G. Withers,
Protein Eng. Des. Sel. 2005, 18, 3 3 .
[23] N. D. Meadow, D. K. Fox, S. Roseman, Annu. Rev. Biochem.
1990, 59, 497.
which cannot cleave unactivated thioglycosides. The efficient
hydrolysis of thiodisaccharides therefore provides further
supporting evidence for the novel glycosidase mechanism
proposed for family 4 enzymes. Cleavage of such thioglyco-
sides at rates comparable to those of their oxygen counter-
parts is quite reasonable for the anionic mechanism proposed
but not for reactions via oxocarbenium-ion-like transition
states. The majority of glycosidases do not hydrolyze thio-
glycosidic linkages, and those that have been reported to
possess thioglycosidase activity only react with thioglycosides
containing highly activated leaving groups, as clearly demon-
strated by the Brønsted plot determined for thioglycoside
hydrolysis in the case of O-GlcNAcase.[21] The cleavage of the
thioglycosidic linkages by O-GlcNAcase does not rely on
general acid catalysis, and there is substantial development of
negative charge on the sulfur atom at the transition state.[21]
Since the cleavage of the glycosidic linkage is not rate-limiting
for BglT,[24] it is reasonable that substitution of the glycosidic
oxygen with a sulfur atom does not significantly affect the
overall rate.
[24] V. L. Y. Yip, S. G. Withers, Biochemistry 2006, 45, 571.
[25] W. P. Burmeister, S. Cottaz, H. Driguez, R. Iori, S. Palmieri, B.
Henrissat, Structure 1997, 5, 663.
On the basis of these results, it is clear that thioglycosides
should not be employed in any inhibition strategies for
family 4 enzymes whether this be for structural or mechanistic
studies or as part of any biological control strategy. Indeed,
the ability to cleave a thiodisaccharide could be a useful
diagnostic of whether a new glycosidase belongs to family 4.
Finally, these findings raise the question of whether O-C6’P is
in fact the natural substrate for BglT, or whether this enzyme
has evolved to cleave some other, as yet undiscovered,
substrate. The locations of the genes encoding many GH4
enzymes within the PEP:PTS operon argues strongly that the
disaccharide 6-phosphates are the natural substrates and that
the facile cleavage of thioglycosides is just a circumstantial
consequence of the mechanism utilized by the enzyme.
However, the uncanny similarities between the mechanisms
utilized by GH4 enzymes and by AdoHCyase suggests that
other possibilities should be kept in mind.
Received: April 10, 2006
Revised: June 20, 2006
Published online: August 17, 2006
Keywords: elimination · enzyme catalysis · glycoside hydrolase ·
.
thioether cleavage · thioglycosides
[1] Z. J. Witczak, Curr. Med. Chem. 1999, 6, 165.
[2] N. Asano, Glycobiology 2003, 13, 93R.
[3] D. Mandelman, A. Belaich, J. P. Belaich, N. Aghajari, H.
Driguez, R. Haser, J. Bacteriol. 2003, 185, 4127.
[4] G. Sulzenbacher, H. Driguez, B. Henrissat, M. Schulein, G. J.
Davies, Biochemistry 1996, 35, 15280.
[5] M. L. Sinnott, Chem. Rev. 1990, 90, 1171.
[6] D. L. Zechel, S. G. Withers, Acc. Chem. Res. 2000, 33, 11.
[7] D. E. Koshland, Biol. Rev. 1953, 28, 416.
[8] J. L. Jensen, W. P. Jencks, J. Am. Chem. Soc. 1979, 101, 1476.
[9] J. P. Ferraz, E. H. Cordes, J. Am. Chem. Soc. 1980, 102, 1488.
[10] T. H. Fife, T. J. Przystas, J. Am. Chem. Soc. 1980, 102, 292.
[11] M. Bertoldi, B. Cellini, T. Clausen, C. B. Voltattorni, Biochem-
istry 2002, 41, 9153.
6182
www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 6179 –6182