The chameleon of retaining glycoside hydrolases and retaining glycosyl transferases: The catalytic nucleophile

Article abstract of DOI:10.1007/s007060200027

This paper addresses the existence and role of a catalytic nucleophile in retaining glycoside hydrolases and retaining glycosyl transferases. Although the former has now been established beyond doubt, such is not the case with the latter. We report reliable procedures for the synthesis of various 2-deoxy-2-fluoro glycosyl nucleoside diphosphates, useful donor analogues for the study of the mechanism of action of retaining glycosyl transferases.

Full text of DOI:10.1007/s007060200027

Monatshefte fuÈr Chemie 133, 541±554 ꢀ2002)
The Chameleon of Retaining Glycoside
Hydrolases and Retaining Glycosyl
Transferases: The Catalytic Nucleophile
Ã
Robert V. Stick and AndrewG. Watts
Department of Chemistry, The University of Western Australia, Crawley WA, 6009 Australia
Summary. This paper addresses the existence and role of a catalytic nucleophile in retaining
glycoside hydrolases and retaining glycosyl transferases. Although the former has now been
established beyond doubt, such is not the case with the latter. We report reliable procedures for the
synthesis of various 2-deoxy-2-¯uoro glycosyl nucleoside diphosphates, useful donor analogues for
the study of the mechanism of action of retaining glycosyl transferases.
Keywords. Glycoside hydrolases; Glycosyl transferases; 2-Deoxy-2-¯uoro hexopyranoses.
Introduction
The burgeoning ®eld of glycobiology arose from the recognition that carbohydrates
play a pivotal role in an overwhelming number of biological processes [1±6].
Carbohydrates are now known to be of key importance in cell adhesion, infection,
differentiation, and regulation as well as in many intercellular communication and
signal transduction events [7]. In order to gain a deeper understanding of these
events, it was natural for chemists, biochemists, and molecular biologists to probe
the mechanism of action of the enzymes that process carbohydrates, mainly the
glycoside hydrolases and glycosyl transferases.
This article, therefore, will ®rst address some recent efforts with glycoside
hydrolases, and then turn to the more vexed question regarding the mechanism of
action of retaining glycosyl transferases.
Results and Discussion
Glycoside hydrolases
The mode of action of glycoside hydrolases, enzymes that catalyze the hydrolytic
cleavage of glycosidic bonds in both simple and complex carbohydrate structures,
is thoroughly understood thanks to the early and seminal efforts of Koshland [8]
and Legler [9], and more recently by Sinnott [10, 11], Vasella [12], Davies [13, 14],
Ã
Corresponding author. E-mail: rvs@chem.uwa.edu.au

542
R. V. Stick and A. G. Watts
and Withers [15±20]. The glycoside hydrolases have been classi®ed into families
based on amino acid sequence homology [13, 21, 22], and most of these families
more broadly into clans [23]. The members of a family all exhibit the same
mechanism of action, whether they be exo- or endo-hydrolases and whether they
act on similar or unrelated substrates. More recently, glycoside hydrolases have
also been classi®ed according to the direction from which the glycosidic oxygen is
protonatated during hydrolysis ꢀsyn- or anti-protonators) [24].
Glycoside hydrolases catalyze the cleavage of the glycosidic linkage with either
inversion or retention of con®guration at the anomeric centre; the well established
steps in the two separate processes are shown below ꢀfor a ꢀ-D-glycoside) [8].
Inversion:
Retention:
For an inverting glycoside hydrolase, general acid-base catalysis prevails, and
there is a need for pKa modulation of the two relevant carboxylic acid residues. The
active site, in order to accommodate the substrate and a molecule of water, is some
Ê
10 A across [13]. The salient points of a retaining glycoside hydrolase are the
Ê
smaller active site ꢀ6 A), the formation of a true glycosyl-enzyme intermediate
through the agency of a catalytic nucleophile ꢀaspartate or glutamate), and the
stabilization of the transition states associated with steps a and b by the hydroxyl
group present at C2 [13, 19].

Retaining Glycoside Hydrolases and Glycosyl Transferases
543
The main ꢀand obvious) difference between an inverting and a retaining
hydrolase is the presence of a carboxylic acid residue capable of acting as a
catalytic nucleophile. This nucleophile was ®rst detected and, in some instances,
identi®ed by the use of af®nity labels such as epoxyalkyl glycosides [9] ± these
molecules somewhat resemble the natural substrate for the enzyme and, addi-
tionally, are capable of ꢀsometimes indiscriminate) attachment to a carboxylic
residue within the enzyme [25, 26].
Much more impressive as labels of the catalytic nucleophile are the deoxy
¯uoro sugars recently developed by Withers. A 2-deoxy-2-¯uoro sugar, suitably
activated at the anomeric carbon with an excellent leaving group, is able to label
the catalytic nucleophile speci®cally [27±29].
The glycosyl-enzyme intermediate is suf®ciently long-lived ꢀowing to the
stabilizing effect of the ¯uorine atom) as to be detected and observed by nuclear
magnetic resonance, mass spectrometric, or X-ray crystallographic techniques.
Nowhere has this approach been put to better use than in the recent labelling of the
catalytic nucleophile of hen egg-white lysozyme, long accepted as the paradigm
for retaining glycoside hydrolases and now shown to operate in common with other
members of the family [30].
For somewhat related reasons, 5-¯uoro sugars, also activated at the anom-
eric carbon, were found to be effective reagents for the labelling of the cata-
lytic nucleophile [18, 31, 32]. The establishment of a mechanism for retaining
glycoside hydrolases has led to effective procedures for glycoside synthesis
through the action of transglycosylation and the invention of glycosynthases
[33, 34].
We were attracted to the synthesis of 2-deoxy-2-¯uoro sugars some years ago,
but sorely missed the boat in terms of publication. However, we still feel that some
comment is warranted here. Although there are various methods available for the
synthesis of 2-deoxy-2-¯uoro sugars, none are general, and some involve the use of

544
R. V. Stick and A. G. Watts
obnoxious reagents such as acetyl hypo¯uorite or elemental ¯uorine [35±37]. The
whole situation changed with the advent of Select¯uorTM, a ¯uoroammonium salt:
Wong was the ®rst to report the addition of Select¯uorTM to glycals, ultimately
to produce mixtures of free sugars and glycosides [38].
The stereoselectivity of the process can be improved by using bulky protecting
groups on the glycal or through the use of a favourable substrate [39, 40].
Most of the mechanistic studies involving the addition of Select¯uorTM to
glycals have been conducted by Dax [39, 41] and Wong [40] and, together
with these results, we propose a rationalization for the process. The addition, by
necessity, is a concerted process, and with a substrate such as tri-O-acetyl-D-glucal,
two glycosylammonium salts result [39].
4
The 2-deoxy-2-¯uoro-D-mannopyranosyl salt would be present in the C1
conformation, and there is some evidence that the corresponding D-glucopyranosyl
salt is also dominated by a reverse anomeric effect and found in the C4 confor-
1
mation [40]. A subsequent hydrolysis, for example, yields the observed mixtures of
free sugars. With glycals such as tri-O-pivaloyl-D-glucal and tri-O-acetyl-D-
galactal, the addition of Select¯uorTM occurs from the less hindered face; di-O-
acetyl-L-fucal is an ideal substrate in several ways.

Retaining Glycoside Hydrolases and Glycosyl Transferases
545
It is also worth mentioning that some glycal derivatives bearing substituents at
C2 are found to be unreactive towards Select¯uorTM. Tri-O-acetyl-1,2-dideoxy-2-
¯uoro-D-arabino-hex-1-enopyranose, when treated with Select¯uorTM for several
days, failed to produce tri-O-acetyl-2-deoxy-2,2-di¯uoro-D-glucopyranose. A sim-
ilar lack of reactivity has been observed with tri-O-acetyl-1,2-dideoxy-2-ꢀN,N-
diacetylamino)-D-arabino-hex-1-enopyranose [42].
Glycosyl transferases
The discussion here will be restricted to those enzymes that catalyze the transfer of
a sugar residue from a nucleoside diphosphate to an acceptor molecule, usually
another carbohydrate.
Investigations into the mechanism of action of such glycosyl transferases have
sorely lagged behind those of the glycoside hydrolases. The reasons for this lag are
numerous, including dif®culties in access and supply and the fact that many of the
glycosyl transferases are membrane associated ± crystallization of the membrane-
free enzyme has been, until recently, a dif®cult affair.
However, in spite of the above problems, glycosyl transferases have been
classi®ed into families ꢀalong lines similar to those for the glycoside hydrolases),
even though the amino acid sequence homology for transferases can be virtually
non-existent [43, 44]. The transferases are also classed as inverting or retaining,
and this characteristic of the enzyme is easily assessed ꢀin contrast to the glycoside
hydrolases) by a simple inspection of the reaction products.
Perhaps one of the follies of carbohydrate chemists and biochemists alike in the
late twentieth century will be the belief that the mechanistic principles established
for glycoside hydrolases could be applied directly, almost without question, to
glycosyl transferases.
Inversion:

546
R. V. Stick and A. G. Watts
Retention:
The application of such mechanistic principles presents few problems for an
inverting glycosyl transferase, with some ꢀlimited) experimental evidence to sup-
port the case [45±47]. However, a retaining enzyme again requires the involvement
of a catalytic nucleophile, generally suggested to be a carboxylic acid ꢀaspartate or
glutamate) residue. Although the presence of a carboxylic catalytic nucleophile is
suggested on the basis of results from non-speci®c labelling reagents ꢀcarbodi-
imides) and amino acid sequence alignments ꢀfor glycogen synthase) [48], there
has never been reported the actual labelling and identi®cation of such a residue for
a glycosyl transferase.
Until recently, no glycosyl transferase had been crystallized, so as to allow
mechanistic investigations along the lines of the beautiful piece of work on
glycoside hydrolases by Davies and Withers [49]. However, in the last decade,
six glycosyl transferases, all inverting in their action, were crystallized and
investigated by X-ray crystallography [50±55]. Several of these enzymes required
the presence of a nucleoside diphosphate donor for a successful outcome [51±53].
Unfortunately, no great insights into the enzymes were revealed, particularly
when the actual identity of some of the natural substrates remained unknown
[52].
The seventh glycosyl transferase to be crystallized, however, was a retaining
enzyme [56]. The successful experiment was conducted with a bacterial 1,4-ꢁ-
galactosyl transferase ꢀfamily 8), modi®ed by point mutations, and crystallized in
the presence of a deactivated donor and modi®ed acceptor:

Retaining Glycoside Hydrolases and Glycosyl Transferases
547
Surprisingly, the authors could ®nd no evidence for the presence of any
catalytic nucleophile. In the absence of a catalytic carboxylic acid ꢀAsp and Glu)
residue, several other contenders were considered, including a glutamine residue
and the 6-OH of the acceptor molecule. In some desperation, a chemically unusual
SNi mechanism was ®nally proposed [56].
The most striking points for us in this very important paper were the absence of
an obvious catalytic nucleophile and the necessity for a metal ion ꢀMn2‡ ) for
successful catalysis. We have been working for some time now on another family 8
glycosyl transferase, glycogenin, a self-glucosylating enzyme that initiates the
synthesis of macro-glycogen [57, 58]:
Glycogenin is an ꢁ-retaining enzyme, and we have sought to label the
catalytic nucleophile by incubation of the enzyme with UDP-2-deoxy-2-¯uoro-ꢁ-
D-glucose ± the results have been, at best, equivocal. However, a notable fact is that
glycogenin is also metal ion ꢀMn2‡ ) dependant [59]. If it were to eventuate that
some retaining glycosyl transferases do not employ a nucleophile in their catalytic
event, what then is their mechanism of action? For family 8 retaining transferases,
one of the few conserved residues ꢀDXD) is doubtlessly involved in the binding of
the metal ion ꢀMn2‡ ). This leads us to suggest a general process for the catalytic
action of ꢀfamily 8) retaining glycosyl transferases:
ꢀ1) The initial event involves binding of the glycosyl donor into the active site,
partly stabilized by an interaction between the pyrophosphate residue and the
metal ion ꢀMn2‡ ).
ꢀ2) Next follows the binding of the acceptor, which triggers a change in the
shape of the enzyme at the active site, ensuring an anhydrous environment
[52, 53].
ꢀ3) This change in conformation also allows a hydrogen bond to form between the
hydroxyl group of the acceptor and the glycosidic oxygen atom, causing
cleavage of the glycosidic linkage and generating an oxacarbenium ion that
may be stabilized through non-bonding interactions with adjacent amino acid
residues.
ꢀ4) The oxacarbenium ion, through the agency of the aforementioned conforma-
tional change, is now adjacent to the acceptor and forms the new glycosidic
linkage from the appropriate face.

548
R. V. Stick and A. G. Watts
ꢀ5) The enzyme undergoes another conformational change, ®rst releasing the new
glycoside, followed by the nucleoside diphosphate, thus presenting the enzyme
in its native form ready to repeat the cycle.
In the case of the bacterial 1,4-ꢀ-galactosyl transferase mentioned previously,
the oxacarbenium ion predicted in ꢀ3) is likely to be stabilized by a highly
conserved glutamine residue ꢀFig. 1).
We feel that some other points are worth mentioning in the context of the above
proposal. With retaining glycoside hydrolases, water is the natural acceptor
molecule for the glycosyl-enzyme intermediate; when such hydrolases are used in
an unnatural process, i.e. glycoside synthesis, the acceptor molecule is generally an
alcohol, much less reactive than water and re¯ected in a slower process. However,
the presence of a general acid=base residue in the hydrolase does allow the
synthetic process to proceed [34]. If one now translates these observations to
retaining glycosyl transferases, hydrolysis of the donor is the unnatural process,
and the natural acceptor is an unreactive alcohol. Surely this situation requires a
donor far more reactive than a conventional glycosyl-enzyme intermediate, es-
pecially when a general acid=base residue appears to be absent. We believe that
the new hydrogen bond formed between the acceptor hydroxyl group and the
glycosidic oxygen of the donor holds the key.
Fig. 1. Depiction of a modi®ed active site arrangement for bacterial 1,4-ꢀ-galactosyl transferase
ꢀPDB: 1GA8) showing bond distances between C1 and Gln189 ꢀrotated) as well as O1 and OH400
ꢀinserted)

Retaining Glycoside Hydrolases and Glycosyl Transferases
549
Syntheses
Our ®nal words in this paper are concerned with the synthesis of the glycosyl
donors ꢀand related molecules) used in the above investigations: 2-deoxy-2-¯uoro
glycosyl nucleoside diphosphates.
For the synthesis of NDP-2-deoxy-2-¯uoro-ꢁ-D-glucose derivatives, tri-O-
acetyl-D-glucal ꢀScheme 1) was treated with Select¯uorTM to give, after selective
acetylation and puri®cation, tetra-O-acetyl-2-deoxy-2-¯uoro-ꢁ-D-glucopyranose.
Inversion of con®guration at the anomeric centre via the agency of the bromide
gave the ꢀ-D-anomer, ready for a successful MacDonald phosphorylation to
provide the ꢁ-D-glucosyl phosphate isolated as the diꢀcyclohexylammonium) salt
[60]. This salt was then converted into the desired nucleoside derivative ꢀUDP or
ADP), employing a modi®ed morpholidate coupling procedure described by
Wittmann and Wong [61]. It should be noted that this transformation could be
conveniently monitored by HPLC, and was found to proceed well using the tri-n-
butylammonium salt of the ꢁ-D-glucopyranosyl phosphate.
Interestingly, the major product from the Select¯uorTM reaction on tri-O-
acetyl-D-glucal, tetra-O-acetyl-2-deoxy-2-¯uoro-ꢁ-D-mannopyranose ꢀScheme 1),
is inert to the MacDonald phosphorylation procedure ± a result of the double
deactivation at the anomeric carbon by an axial acetyl group and an adjacent
¯uorine atom [42]. Instead, tetra-O-acetyl-2-deoxy-2-¯uoro-ꢁ-D-mannopyranose
was converted via the bromide exclusively to the ꢁ-D-anomer of the hemiacetal
ꢀScheme 2). Treatment of the hemiacetal with diphenyl chlorophosphate then gave
the diphenyl phosphate that was convertible, after catalytic hydrogenolysis, into the
desired GDP-2-deoxy-2-¯uoro-ꢁ-D-mannopyranose.
Scheme 1

550
R. V. Stick and A. G. Watts
Scheme 2
Conclusions
There is no doubt that two great papers have appeared in the ®rst year of the new
century ± the ®rst, on the mechanism of action of hen egg-white lysozyme, all but
closes the chapter on retaining glycoside hydrolases; the second, on the search for a
catalytic nucleophile in a retaining glycosyl transferase, answers some questions
and asks of others.
Much work remains to be done, and an upcoming report on the X-ray structure
determination of glycogenin [62], another retaining glycosyl transferase of family
8, is sure to continue the debate1.
Experimental
Experimental details have been given previously [64]. Bio-Gel P-2 Gel ꢀFine) was purchased from
Bio-Rad Laboratories.
Tetra-O-acetyl-2-deoxy-2-¯uoro-ꢁ-D-glucopyranose and Tetra-O-acetyl-
2-deoxy-2-¯uoro-ꢁ-D-mannopyranose ꢀC14H19FO9)
Freshly powdered Select¯uorTM ꢀ5.7 g, 16 mmol) was added to a solution of tri-O-acetyl-D-glucal
ꢀ2.2 g, 8.0 mmol) in H2O=DMF ꢀ100 cm3, 1:1), and the mixture was left to stir at 40ꢀ ꢀ8 h). The
resultant solution was concentrated in vacuo and then subjected to a usual workup ꢀEtOAc) to give a
colourless syrup. Ac2O ꢀ10 cm3) and I2 ꢀ300 mg) were then added to this syrup, and the mixture was
1
During the preparation of this manuscript, a report on a retaining 1,3-ꢁ-galactosyl transferase by
Gastinel and coworkers suggested the detection of a glycosyl-enzyme intermediate [63]; without
access to the primary crystallographic data ꢀelectron density maps), the claim appears somewhat
equivocal.

Retaining Glycoside Hydrolases and Glycosyl Transferases
551
allowed to stand at room temperature ꢀ3 h). The mixture was then concentrated in vacuo, subjected to
a usual workup ꢀEtOAc), and puri®ed using ¯ash chromatography ꢀ30% EtOAc=petroleum ether).
First to elute was the D-gluco derivative ꢀ748 mg, 28%); m.p.: 75±77ꢀC ꢀRef. [65]: 78ꢀC),
[ꢁ]D ˆ ‡ 147ꢀ ꢀRef. [65]: ‡ 148ꢀ); 1H NMR ꢀ300 MHz, CDCl3): ꢂ ˆ 2.01, 2.15, 2.22, 2.27 ꢀ4s, 12H,
CH3), 3.96±4.02 ꢀm, H5,6), 4.23 ꢀdd, J5,6 ˆ 5.1, J6,6 ˆ 11.5 Hz, H6), 4.61 ꢀddd, J2,F ˆ 47.6, J1,2 ˆ 4.7,
J2,3 ˆ 9.5 Hz, H2), 5.02 ꢀt, J3,4 ꢁ J4,5 ˆ 9.6 Hz, H4), 5.48 ꢀdt, J3,F ˆ 13.1 Hz, H3), 6.35 ꢀd, H1) ppm.
Next to elute was the D-manno derivative ꢀ1.34 g, 50%); m.p.: 68±69ꢀC ꢀRef. [35]: 66±68ꢀC),
1
[ꢁ]D ˆ ‡ 65ꢀ ꢀRef. [35]: ‡ 60ꢀ); H NMR ꢀ300 MHz, CDCl3): ꢂ ˆ 2.12, 2.15, 2.22, 2.27 ꢀ4s, 12H,
CH3), 4.13 ꢀddd, J4,5 ˆ 9.6, J5,6 ˆ 5.1, 3.1 Hz, H5), 4.15 ꢀdd, J6,6 ˆ 11.9 Hz, H6), 6.28 ꢀdd, H6), 4.73
ꢀdt, J2,F ˆ 48.9, J1,2 ꢁ J2,3 ˆ 2.1 Hz, H2), 5.23 ꢀddd, J3,F ˆ 23.0, J3,4 ˆ 9.7 Hz, H3), 5.41 ꢀdd, H4),
6.23 ꢀJ1,F ˆ 8.5 Hz, H1) ppm.
Di,cyclohexylammonium) 2-deoxy-2-¯uoro-ꢁ-D-glucopyranosyl
phosphate ꢀC18H38FN2O8P)
Tetra-O-acetyl-2-deoxy-2-¯uoro-ꢀ-D-glucopyranose [60] ꢀ1.15 g) was stirred with molten anhydrous
H3PO4 ꢀ4.0 g) under vacuum ꢀ1 mm Hg) at 60ꢀC overnight. The mixture was allowed to cool slightly
ꢀbut not enough to allow crystallization) and then treated with THF ꢀ5 cm3) to give a dark solution.
Aqueous 2 M LiOH ꢀ80 cm3) was then added to the solution with cooling, and the mixture was left to
stir at room temperature overnight. The mixture was ®ltered and the ®ltrate passed through a column
‡
of Dowex 50W-X8 ꢀH ) resin into an excess of cyclohexylamine in H2O. Concentration of the
eluent in vacuo and subsequent lyophilization gave diꢀcyclohexylammonium) 2-deoxy-2-¯uoro-
ꢁ-D-glucopyranosyl phosphate as a pale yellow powder ꢀ1.32 g, 87%). [ꢁ]D ˆ ‡46ꢀ ꢀRef. [60]:
1
‡ 48.6ꢀ); the H NMR spectrum was consistent with that reported [60].
Tri-O-acetyl-2-deoxy-2-¯uoro-ꢁ-D-mannopyranose ꢀC12H17FO8)
HBr in AcOH ꢀ2 cm3, 30% w=w) was added to tetra-O-acetyl-2-deoxy-2-¯uoro-ꢁ-D-mannopyranose
ꢀ1.80 g, 5.4 mmol), and the mixture was left to stir at room temperature ꢀ4 h). Usual workup ꢀEtOAc)
gave the ꢁ-D-mannosyl bromide as a colourless oil ꢀ2.1 g) which was used without further
puri®cation.
Ag2CO3 ꢀ1.14 g, 4.13 mmol) was added to a solution of the above bromide ꢀ1.4 g, 3.8 mmol) in
aqueous acetone ꢀ20 cm3, 1:10), and the mixture left to stir ꢀ2 h). The mixture was then ®ltered
through a plug of silica and concentrated in vacuo. Usual workup ꢀEtOAc) of the residue followed by
¯ash chromatography ꢀ40% EtOAc=petroleum ether) gave the title hemiacetal as a colourless oil
ꢀ1.15 g, 98%). The spectroscopic data ꢀ1H and 13C NMR) were consistent with those reported [39].
Diphenyl ,tri-O-acetyl-2-deoxy-2-¯uoro-ꢁ-D-mannopyranosyl) phosphate ꢀC24H26FO11P)
A solution of tri-O-acetyl-2-deoxy-2-¯uoro-ꢁ-D-mannopyranose ꢀ0.76 g, 2.5 mmol) and DMAP
ꢀ0.55 g) in CH2Cl2 ꢀ20 cm3) was cooled to À10ꢀC. Diphenyl chlorophosphate ꢀ1.0 cm3, 4.8 mmol)
was added dropwise, and the solution was allowed to stir ꢀ4 h) under gradual warming to 10ꢀC. The
solution was diluted with CH2Cl2 ꢀ20 cm3) and washed sequentially with cold H2O, 0.5 M HCl, and
saturated NaHCO3 solution, dried ꢀMgSO4), and concentrated in vacuo. Flash chromatography ꢀ20%
EtOAc=petroleum ether) of the residue gave the title phosphate as a colourless oil ꢀ1.01 g, 76%).
[ꢁ]D ˆ ‡ 46ꢀ; found: C 53.5, H 4.7; calcd.: C 53.3, H 4.8; 1H NMR ꢀ300 MHz, CDCl3): ꢂ ˆ 1.95,
2.00, 2.09 ꢀ3s, 9H, CH3), 3.88 ꢀdd, J5,6 ˆ 1.4, J6,6 ˆ 12.5 Hz, H6), 3.98±4.04 ꢀm, H5), 4.13 ꢀdd,
J5,6 ˆ 4.2 Hz, H6), 4.74 ꢀdddd, J2,F ˆ 48.9, J1,2 ꢁ J2,3 ˆ 2.2, J2,P ˆ 0.5 Hz, H2), 5.25 ꢀddd, J3,F ˆ 27.2,
J3,4 ˆ 10.3 Hz, H3), 5.37 ꢀddd, J4,F ˆ 1.2, J4,5 ˆ 10.2 Hz, H4), 5.99 ꢀddd, J1,F, J1,P ˆ 7.1, 8.5 Hz, H1),
7.15±7.39 ꢀm, 10H, Ph) ppm; 13C NMR ꢀ75.5 MHz, CDCl3): ꢂ ˆ 20.47, 20.55 ꢀ3C, CH3), 61.19 ꢀC6),

552
R. V. Stick and A. G. Watts
64.59, 70.67 ꢀC4,5), 68.75 ꢀd, J3,F ˆ 17 Hz, C3), 85.98 ꢀdd, J2,F ˆ 182, J2,P ˆ 12 Hz, C2), 95.19 ꢀdd,
J1,F, J1,P ˆ 6.0, 32 Hz, Cl), 119.90±150.11 ꢀ12C, Ph) ppm.
2-Deoxy-2-¯uoro-ꢁ-D-mannopyranosyl-bis,tributylammonium) phosphate ꢀC30H66FN2O8P)
PtO2 ꢀ40 mg) was added to a solution of diphenyl 3,4,6-tri-O-acetyl-2-deoxy-2-¯uoro-ꢁ-D-mannosyl
phosphate ꢀ400 mg) in MeOH ꢀ15 cm3), and the mixture was stirred vigorously under an atmosphere
of H2 at room temperature ꢀ2 d). The mixture was ®ltered, diluted with H2O ꢀ2 cm3) and Et3N
ꢀ2 cm3), and the solution was left to stand at 4ꢀC overnight. The mixture was concentrated in vacuo,
‡
taken up in H2O ꢀ5 cm3), and passed through a column of Dowex 50 W-X8 ꢀH ) resin into a vig-
orously stirred mixture of H2O and excess tributylamine. Concentration in vacuo and subsequent
lyophilization gave the title phosphate as a white powder ꢀ416 mg).
1H NMR ꢀ300 MHz, D2O): ꢂ ˆ 0.72 ꢀt, 18H, CH3), 1.10±1.21, 1.35±1.51 ꢀ2m, 24H, CH2), 2.85±
2.98 ꢀm, 12H, CH2N), 3.48±3.69 ꢀm, 4H, H4,5,6), 3.76 ꢀddd, J3,F ˆ 30.8, J2,3 ˆ 2.5, J3,4 ˆ 9.6 Hz,
H3), 4.59 ꢀbddd, J2,F ˆ 48.7, J1,2 ˆ 2.4 Hz, H2), 5.38 ꢀddd, J1,P, J1,F ˆ 6.0, 8.1 Hz, H1) ppm; 13C
NMR ꢀ75.5 MHz, D2O): ꢂ ˆ 24.05 ꢀ6C, CH3), 30.52, 36.66 ꢀ12C, CH2), 63.89 ꢀ6C, CH2N), 71.57
ꢀC6), 77.62, 84.66 ꢀC4,5), 80.40 ꢀd, J3,F ˆ 17.3 Hz, C3), 100.97 ꢀdd, J2,F ˆ 174.4, J2,P ˆ 9.2 Hz, C2),
104.18 ꢀdd, J1,F ˆ 4.5, J1,P ˆ 30.6 Hz, C1) ppm.
Uridine 50-,trihydrogendiphosphate), P0-,2-deoxy-2-¯uoro-ꢁ-D-glucopyranosyl) ester
ꢀC15H23FN2O16P2)
Diꢀcyclohexylammonium) 2-deoxy-2-¯uoro-ꢁ-D-glucopyranosyl phosphate ꢀ370 mg, 803 ꢃmol)
‡
was dissolved in H2O ꢀ5 cm3) and passed through a column of Dowex 50 W-X8 ꢀH ) resin into a
vigorously stirred mixture of H2O and excess tributylamine. The eluent was then concentrated
in vacuo, coevaporated with MeOH ꢀ2Â 10 cm3), and dried under high vacuum ꢀ2 d) to give the
bisꢀtributylammonium) salt ꢀ410 mg).
1H NMR ꢀ300 MHz, D2O): ꢂ ˆ 0.72 ꢀt, 18H, CH3), 1.07±1.16, 1.25±1.37 ꢀ2m, 24H, CH2), 2.89±
2.93 ꢀm, 12H, CH2N), 3.29 ꢀt, J3,4 ꢁ J4,5 ˆ 9.9 Hz, H4), 3.56 ꢀdd, J5,6 ˆ 4.8, J6,6 ˆ 12.3 Hz, H6),
3.63±3.69 ꢀm, H5,6), 3.80 ꢀdt, J2,3 ˆ 9.6, J3,F ˆ 13.1 Hz, H3), 4.21 ꢀdddd, J1,2, J1,P ˆ 2.3, 3.7,
J2,F ˆ 49.4 Hz, H2), 6.13 ꢀdd, J1,P ˆ 6.6 Hz, H1) ppm; 13C NMR ꢀ125.8 MHz, D2O): ꢂ ˆ 15.14 ꢀ6C,
CH3), 21.61, 27.51 ꢀ12C, CH2), 54.96 ꢀ6C, CH2N), 62.52 ꢀC6), 71.15 ꢀd, J4,F ˆ 8.3 Hz, C4), 73.37 ꢀd,
J3,F ˆ 17.3 Hz, C3), 74.78 ꢀC5), 91.89 ꢀdd, J2,P ˆ 6.8, J2,F ˆ 187 Hz, C2), 94.25 ꢀdd, J1,P ˆ 4.6,
J1,F ˆ 22.3 Hz, C1) ppm.
The above bisꢀtributylammonium) salt ꢀ410 mg) and 4-morpholine-N,N0-dicyclohexylcarbox-
amidinium uridine 50-monophosphomorpholidate ꢀ662 mg, 964 ꢃmol) were coevaporated with dry
pyridine ꢀ3 Â 2 cm3). 1H-Tetrazole ꢀ112 mg, 1.61 mmol) and dry pyridine ꢀ4 cm3) were added, and
the solution was then concentrated to half of its original volume and left to stir at room temperature
ꢀ2 d). The solution was then diluted with H2O ꢀ30 cm3), washed with Et2O, and concentrated
in vacuo. The residue was puri®ed on a Bio-Gel P-2 column ꢀ2.5 Â 95 cm) eluting with 250 mM
NH4HCO3 solution and lyophilized to give the title compound as a white powder ꢀ320 mg, 68%).
1H NMR ꢀ500 MHz, D2O): ꢂ ˆ 3.34 ꢀt, J3 ;4 ꢁ J4 ;5 ˆ 9:8 Hz, H400), 3.59 ꢀdd, J5 ;6 ˆ 4:5,
00 00
00 00
00 00
J6 ;6 ˆ 12:5 Hz, H600), 3.67 ꢀdd, J5 ;6 ˆ 2:2 Hz, H600), 3.73 ꢀddd, H500), 3.85 ꢀdt, J2 ;3 ˆ 9:7, J3 ;F
ˆ
00 00
00 00
00 00
00
12:8 Hz, H300), 3.97±4.10, 4.16±4.22, 4.27±4.32 ꢀ3m, 6H, H20,30,40,50,200), 5.59 ꢀdd, J1 ;2 ˆ 3:5,
00 00
J1 ;F ˆ 7:3 Hz, H100), 5.77±5.82 ꢀm, H5,10), 7.78 ꢀd, J5,6 ˆ 8.1 Hz, H6) ppm.
00
Adenosine 50-,trihydrogendiphosphate), P0-,2-deoxy-2-¯uoro-ꢁ-D-glucopyranosyl)
ester ꢀC16H24FN5O14P2)
Following the above procedure, diꢀcyclohexylammonium) 2-deoxy-2-¯uoro-ꢁ-D-glucopyranosyl
phosphate ꢀ285 mg, 618 ꢃmol) was treated with 4-morpholine-N,N0-dicyclohexylcarboxamidinium

Retaining Glycoside Hydrolases and Glycosyl Transferases
553
adenosine 50-monophosphomorpholidate ꢀ527 mg, 742 ꢃmol) and 1H-tetrazole ꢀ88 mg, 1.2 mmol) in
dry pyridine ꢀ2 cm3) to give the title compound as a hygroscopic white solid ꢀ257 mg, 71%).
1H NMR ꢀ300 MHz, D2O): ꢂ ˆ 3.35 ꢀdd, J3 ;4 ˆ 11:1, J4 ;5 ˆ 10:9 Hz, H400), 3.56±3.78 ꢀm, 3H,
00 00
00 00
H500,600), 3.87 ꢀddd, J2 ,3 ˆ 11.0, J3 ,F ˆ 50.9 Hz, H300), 4.06±4.41 ꢀm, 6H, H20,30,40,50,200), 5.51 ꢀdd,
00 00
00
J1 ,2 ˆ 4.1, J1 ,F ˆ 7.2 Hz, H100), 5.80 ꢀd, J1 ,2 ˆ 4.5 Hz, H1 ), 7.86, 8.00 ꢀ2s, H2,8) ppm.
0
00 00
00
0
0
Guanosine 50-,trihydrogendiphosphate), P0-,2-deoxy-2-¯uoro-ꢁ-D-mannopyranosyl)
ester ꢀC16H24FN5O15P2)
Following the previous procedure, bisꢀtributylammonium) 2-deoxy-2-¯uoro-ꢁ-D-mannopyranosyl
phosphate ꢀ320 mg, 694 ꢃmol) was treated with 4-morpholine-N,N0-dicyclohexylcarboxamidinium
guanosine 50-monophosphomorpholidate ꢀ605 mg, 833 ꢃmol) and 1H-tetrazole ꢀ98 mg, 1.4 mmol) in
dry pyridine ꢀ2 cm3) to give the title compound as a white powder ꢀ281 mg, 65%). Spectroscopic data
ꢀ1H and 13C NMR) were consistent with those reported [66].
Acknowledgments
The authors would like to thank the Australian Research Council for ®nancial assistance, as well as
Air Products and Chemicals, Inc. for a kind donation of Select¯uorTM and ®nancial assistance.
References
[1] Wong C-H, Halcomb RL, Ichikawa Y, Kajimoto T ꢀ1995) Angew Chem Int Ed Engl 34: 521
[2] Wong C-H, Halcomb RL, Ichikawa Y, Kajimoto T ꢀ1995) Angew Chem Int Ed Engl 34: 412
[3] Kobata A ꢀ1993) Acc Chem Res 26: 319
[4] Dwek RA ꢀ1995) Biochem Soc Trans 23: 1
[5] Dwek RA ꢀ1996) Chem Rev 96: 683
[6] Sears P, Wong C-H ꢀ1998) Chem Commun 1161
[7] Varki A ꢀ1993) Glycobiology 3: 97
[8] Koshland DE ꢀ1953) Biol RevCambridge Philos Soc 28: 416
[9] Legler G, Bause E ꢀ1973) Carbohydr Res 28: 45
[10] Sinnott ML ꢀ1990) Chem Rev 90: 1171
[11] Sinnott ML ꢀ1993) Bioorg Chem 21: 34
[12] Varrot A, Schulein M, Pipelier M, Vasella A, Davies GJ ꢀ1999) J Am Chem Soc 121: 2621
[13] Davies GJ, Henrissat B ꢀ1995) Structure 3: 853
[14] Davies GJ, Wilson KS, Henrissat B ꢀ1997) Biochem J 321: 557
[15] Withers SG, Street IP, Bird P, Dolphin DH ꢀ1987) J Am Chem Soc 109: 7530
[16] Withers SG, Street IP ꢀ1988) J Am Chem Soc 110: 8551
[17] Withers SG, Rupitz K, Street IP ꢀ1988) J Biol Chem 263: 7929
[18] McCarter JD, Withers SG ꢀ1996) J Am Chem Soc 118: 241
[19] McCarter JD, Withers SG ꢀ1994) Curr Opin Struct Biol 4: 885
[20] McCarter JD ꢀ1995) PhD Thesis, The University of British Columbia
[21] Henrissat B ꢀ1991) Biochem J 280: 309
[22] Henrissat B, Bairoch A ꢀ1993) Biochem J 293: 781
[23] Henrissat B, Bairoch A ꢀ1996) Biochem J 316: 695
[24] Heightman TD, Vasella AT ꢀ1999) Angew Chem Int Ed 38: 750
[25] Gebler JC, Aebersold R, Withers SG ꢀ1992) J Biol Chem 267: 11126
[26] Miao S, McCarter JD, Grace ME, Grabowski GA, Aebersold R, Withers SG ꢀ1994) J Biol Chem
269: 10975
[27] Street IP, Kempton JB, Withers SG ꢀ1992) Biochemistry 31: 9970
[28] Braun C, Brayer GD, Withers SG ꢀ1995) J Biol Chem 270: 26778

554
R. V. Stick and A. G. Watts: Retaining Glycoside Hydrolases and Glycosyl Transferases
[29] Withers SG, Aebersold R ꢀ1995) Prot Sci 4: 361
[30] Vocadlo DJ, Davies GJ, Laine R, Withers SG ꢀ2001) Nature 412: 835
[31] Vocadlo DJ, Meyer C, He S, Withers SG ꢀ2000) Biochemistry 39: 117
[32] Numao S, He S, Evjen G, Howard S, Tollersrud OK, Withers SG ꢀ2000) FEBS Letters 484: 175
[33] Mackenzie LF, Wang Q, Warren RAJ, Withers SG ꢀ1998) J Am Chem Soc 120: 5583
[34] Mayer C, Zechel DL, Reid SP, Warren RAJ, Withers SG ꢀ2000) FEBS Letters 466: 40
[35] Adamson J, Foster AB, Hall LD, Johnson RN, Hesse RH ꢀ1970) Carbohydr Res 15: 351
[36] Adamson J, Foster AB, Westwood JH ꢀ1971) Carbohydr Res 18: 345
[37] Diksic M, Jolly D ꢀ1986) Carbohydr Res 153: 17
[38] Burkart MD, Zhang Z, Hung S-C, Wong C-H ꢀ1997) J Am Chem Soc 119: 11743
[39] Ortner J, Albert M, Weber H, Dax K ꢀ1999) J Carbohydr Chem 18: 297
[40] Vincent SP, Burkart MD, Tsai C-Y, Zhang Z, Wong C-H ꢀ1999) J Org Chem 64: 5264
[41] Albert M, Dax K, Ortner J ꢀ1998) Tetrahedron 54: 4839
[42] Stick RV, Watts AG ꢀunpublished results)
[43] Campbell JA, Davies GJ, Bulone V, Henrissat B ꢀ1997) Biochem J 326: 929
[44] Geremia RA, Petroni EA, Ielpi L, Henrissat B ꢀ1996) Biochem J 318: 133
[45] Qiao L, Murray BW, Shimazaki M, Schultz J, Wong C-H ꢀ1996) J Am Chem Soc 118: 7653
[46] Murray BW, Wittmann V, Burkart MD, Hung S-C, Wong C-H ꢀ1997) Biochemistry 36: 823
[47] Murray BW, Takayama S, Schultz J, Wong C-H ꢀ1996) Biochemistry 35: 11183
[48] Cid E, Gomis RR, Geremia RA, Guinovart JJ, Ferrer JC ꢀ2000) J Biol Chem 275: 33614

[49] Davies GJ, Mackenzie L, Varrot A, Dauter M, Brzozowski AM, Schulein M, Withers SG ꢀ1998)
Biochemistry 37: 11707
[50] Vrielink A, Ruger W, Driessen HPC, Freemont PS ꢀ1994) EMBO J 13: 3413
[51] Gastinel LN, Cambillau C, Bourne Y ꢀ1999) EMBO J 18: 3546
[52] Charnock SJ, Davies GJ ꢀ1999) Biochemistry 38: 6380

[53] Unligil UM, Zhou S, Yuwaraj S, Sarkar M, Schachter H, Rini JM ꢀ2000) EMBO J 19: 5269
[54] Ha S, Walker D, Shi Y, Walker S ꢀ2000) Protein Sci 9: 1045
[55] Pedersen LC, Tsuchid K, Kitagawa H, Sugahara K, Darden TA, Negishi M ꢀ2000) J Biol Chem
275: 34580
[56] Persson K, Ly HD, Dieckelmann M, Wakarchuk WW, Withers SG, Strynadka NCJ ꢀ2001) Nature
Struct Biol 8: 166
[57] Lin A, Mu J, Yang J, Roach PJ ꢀ1999) Arch Biochem Biophys 363: 163
[58] Alonso MD, Lomako J, Lomako WM, Whelan WJ ꢀ1995) FASEB J 9: 1126
[59] Lomako J, Lomako WM, Whelan WJ ꢀ1990) BioFactors 2: 251
[60] Withers SG, MacLennan DJ, Street IP ꢀ1986) Carbohydr Res 154: 127
[61] Wittmann V, Wong C-H ꢀ1997) J Org Chem 62: 2144
[62] Roach PJ ꢀunpublished results)
[63] Gastinel LN, Bignon C, Misra AK, Hindsgaul O, Shaper JH, Joziasse DH ꢀ2001) EMBO J 20: 638
[64] McAuliffe JC, Stick RV ꢀ1997) Aust J Chem 50: 193
[65] Adam MJ ꢀ1982) J Chem Soc Chem Commun 730
[66] Wong C-H ꢀ1998) Inhibitors of Glycosyl Transferases, WO 98=25940 ꢀUSA), 18th June 1998
Received October 16, 2001. Accepted November 7, 2001