Chemo-enzymatic synthesis of fluorinated sugar nucleotide: Useful mechanistic Probes for glycosyltransferases

Article abstract of DOI:10.1016/S0968-0896(00)00139-5

An effective procedure for the synthesis of 2-deoxy-2-fluoro-sugar nucleotides via Selectfluor-mediated electrophilic fluorination of glycals with concurrent nucleophilic addition or chemo-enzymatic transformation has been developed, and the fluorinated sugar nucleotides have been used as probes for glycosyltransferases, including fucosyltransferase III, V, VI, and VII, and sialyl transferases. In general, these fluorinated sugar nucleotides act as competitive inhibitors versus sugar nucleotide substrates and form a tight complex with the glycosyltransferase. Copyright (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0968-0896(00)00139-5

Bioorganic & Medicinal Chemistry 8 (2000) 1937±1946
Chemo-Enzymatic Synthesis of Fluorinated Sugar Nucleotide:
Useful Mechanistic Probes for Glycosyltransferases
Michael D. Burkart,a Stephane P. Vincent,a Arno Du€els,a,b Brion W. Murray,
Steven V. Leyb and Chi-Huey Wonga,*
aDepartment of Chemistry and The Skaggs Institute for Chemical Biology, The Scripps Research Institute,
10550 N. Torrey Pines Road, La Jolla, CA 92037, USA
bDepartment of Chemistry, University of Cambridge, Lens®eld Road, Cambridge CB2 1EW, UK
Received 14 April 2000
AbstractÐAn e€ective procedure for the synthesis of 2-deoxy-2-¯uoro-sugar nucleotides via Select¯uor-mediated electrophilic
¯uorination of glycals with concurrent nucleophilic addition or chemo-enzymatic transformation has been developed, and the
¯uorinated sugar nucleotides have been used as probes for glycosyltransferases, including fucosyltransferase III, V, VI, and VII,
and sialyl transferases. In general, these ¯uorinated sugar nucleotides act as competitive inhibitors versus sugar nucleotide substrates
and form a tight complex with the glycosyltransferase. # 2000 Elsevier Science Ltd. All rights reserved.
Carbohydrates displayed on the cell surface mediate
important molecular recognition processes including bac-
terial and viral infections, cell adhesion in in¯ammation
and cancer metastasis, immune response, di€erentiation,
fertilization and development.1 These complex oligo-
saccharide structures are assembled through the action
of a number of enzymes, predominately by a family of
at least 100 glycosyltransferases which transfer the sugar
moiety from an activated sugar nucleotide donor to an
acceptor sugar.2,3 Selective inhibition of these enzymes
has received increasing attention for the promise of new
therapeutics; however, the e€ort has been met with only
limited success, and continuing studies are necessary for
an understanding of their mechanisms before selective
therapeutics can be developed. Among the many
mechanistic probes developed for the study of glycosyl-
transfer enzymes, ¯uorinated carbohydrates represent
the most commonly used.4
diphosphate synthase. Fluorinated substrate analogues
have also been utilized in the study of carbohydrate-mod-
ifying enzymes.6,7 The proposed transition-state structure
of most carbohydrate-modifying enzymes, including gly-
cosyltransferases and glycosidases, contains a ¯attened
half-chair conformation with substantial oxocarbenium-
ion character at the anomeric position of the glycon
donor.4,8,9 With regard to glycosidases, considerable
structural, kinetic, and mechanistic data have been used
to validate a general transition-state structure that con-
tains discrete oxocarbenium-ion characteristics4,10,11
and ¯uoroglycosides have been used e€ectively to probe
the mechanism of these enzymes.12 17 Replacement of the
2-OH or 5-H group of glycosyl analogues with the strong
electron-withdrawing group ¯uorine transforms the parent
glycoside into a slow substrate for retaining glycosidases,
supporting the rate-determining exo-cyclic glycosidic
cleavage mechanism. These compounds have been used as
mechanism-based inactivators of retaining glycosidases
that form a covalent adduct in the enzyme active site.
Fluorinated substrate analogues have shown historical
utility in probing enzyme mechanism. In the studies of
the prenyl transfer reaction with ¯uorinated allylic iso-
prenoid substrates,5 the linear free energy results
between model solvolysis reactions which proceed
through established carbocationic intermediates and
enzymatic reactions demonstrated a stepwise mechanism
with carbocationic intermediate formation with farnesyl
A similar oxocarbenium-ion transition-state structure has
been supported for several glycosyltransferanes. It has
been shown that b-1,4-galactosyltransferase (b1,4-GalT)
displays a secondary isotope e€ect with UDP-[1-2H]-
galactose (DV=1.21, DV/K=1.05)7 and that uridine 50-
diphospho-2-deoxy-2-¯uoro-a-d-galactose (UDP-2F-Gal)
is a competitive inhibitor of this enzyme (Ki=2mM).18,19
More recently, GDP-[1-2H]-fucose was shown to display
a secondary isotope e€ect with a-1,3-fucosyltransferase V
*Corresponding author. Tel.: +1-858-784-2487; fax: +1-858-784-
2409; e-mail: adrienne@scripps.edu
0968-0896/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(00)00139-5

1938
M. D. Burkart et al. / Bioorg. Med. Chem. 8 (2000) 1937±1946
(FucT-V) (DV=1.26, DV/K=1.22) and that guanosine 50-
diphospho-2-deoxy-2-¯uoro-b-l-fucose (GDP-2F-Fuc)
serves as a competitive inhibitor against this enzyme
(Ki=4.2 mM).8 The observation that iminocyclitols
designed to mimic the charge distribution of the glyco-
sylcation serve as glycosyltransferase inhibitors and a
demonstration of synergistic inhibition by the combination
of an iminocyclitol, GDP, and the acceptor sugar to
mimic the transition-state structure of a fucosyl-
transferase reaction further support this mechanistic
rationale.9,18,20,21 However, the lack of structural data for
glycosyltransferases has made mechanistic conclusions as
well as structure-based inhibitor design very dicult.
Although several acceptor-modi®ed glycosyltransferase
inhibitors have been shown to be excellent inhibitors of
glycosyltransferases, such studies enlightened the
nucleophilic role of the acceptor but have not helped
elucidate the mechanism of nucleotide displacement and
inversion of stereochemistry seen at the donor.22,23
TMS (dH=0.00). 13C NMR spectra are proton decoupled
and the multiplicity of the signals is `s' unless otherwise
noted. 13C chemical shifts are referenced to the solvent
signal (CDCl3, dC=77.0; DMSO-d6, dC=39.5) or to
1,4-dioxane (dC=67.6 in D2O) as internal standard. 19
F
NMR spectra were recorded at 376.5 MHz (Bruker
AMX-400) and referenced to CFCl3 (dF=0.00) as
internal (CDCl3) or external (D2O) standard. 31P NMR
spectra were recorded at 162.0 MHz (Bruker AMX-400)
and referenced to 85% H3PO4 (dp=0.00) as external
standard. Coupling constants were measured in Hertz
(Hz). High resolution mass spectra (HRMS) were
recorded using fast atom bombardment (FAB) method
in a m-nitrobenzyl alcohol matrix doped with NaI or
CsI. Column chromatography was performed on Mal-
linckrodt silica gel 60 (230±400 mesh). Analytical thin-
layer chromatography was performed using silica gel 60
F254 pre-coated glass plates (Merck) and visualized by
quenching of ¯uorescence and (or) by charring after
treatment with cerium molybdophoshate. Size exclusion
chromatography was performed on Bio-Gel P2 Gel
(Bio-Rad Laboratories). Diethyl ether and benzene were
distilled from sodium-benzophenone; dichioromethane,
acetonitrile, and nitromethane from calcium hydride.
Anhydrous pyridine was purchased from Aldrich and
used without further puri®cation. Guanosine 50-diphos-
pho-[U-14C]-b-l-fucose, cytidine 50-monophospho-N-
acetyl[4,5,6,7,8,9-14C]neuraminic acid, and uridine 50-
diphospho-[U-14C]-a-d-galactose were purchased from
Amersham Life Science. ScintiVerse I scintillation
Studies with ¯uorinated disaccharides used to probe the
glycosidase mechanism have deduced that ¯uorine at any
position along the non-reducing sugar serves to impart
mechanism-based inactivation, inhibition, or slow reac-
tivity. This phenomenon grows more pronounced as the
¯uorine substituent nears the anomeric position, making
20-deoxy-10-¯uoro- and 50-¯uoro disaccharides the most
potent inhibitors.4,12,15,16,24,25 We reasoned that ¯ouri-
nated sugar nucleotides would demonstrate greater stabi-
lity and more pronounced e€ects when substituted at the
2- or 6-position of the glycon, as supported by the com-
petitive inhibition against b1,4-GalT by UDP-2F-Gal
and FucT-V by GDP-2F-Fuc.9,19 All these ¯uorinated
sugar derivatives were synthesized through a complex
multi-step procedure.
ꢀ
cocktail and MnCl2 4 H2O were purchased from Fisher
Scienti®c Company. Protein concentrations were deter-
mined with the Coomassie protein staining reagent with
albumin standards as purchased from Pierce. LacNAc-
b-O-(CH2)5 CO2CH3 was prepared according to pub-
lished procedures.26 All other chemicals were purchased
from Aldrich. The scintillation counter used was the
Beckman LS 3801.
Here we describe the e€ective synthesis of ¯uorinated
sugars and their nucleotides using SelectFluor-mediated
reaction with glycals coupled with concurrent chemo-
enzymatic transformations. We also study the e€ect of
these ¯uorinated derivatives of nucleotide sugars on the
fucosyltransferase family (III, V, VI, and VII), sialyl-
transferases (a-2,6 and a-2,3), and galactosyltransferases
(a-1,3 and b-1,4). The synthesis of guanosine 50-diphos-
pho-6-deoxy-6-¯uoro-b-l-fucose (GDP-6F-Fucose) (1)
and GDP-2-deoxy-2-¯uoro-a-l-fucose (8), new ¯uori-
nated GDP-fucose analogues, and a one-pot chemo-enzy-
matic synthesis of UDP-2-deoxy-2-¯uoro-a-d-galactose
(13) are described.
Synthesis of guanosine 50 -diphospho-6-¯uoro-ꢀ-L-fucopyr-
anoside (GDP-6F-Fuc) (1) (Scheme 1):
(A) 1,2,3,4-Di-O-isopropylidene-L-galactopyranoside. An-
hydrous CuSO4 (886 mg, 5.55 mmol) (dried at 140 ꢁC
under vacuum for 18 h) and anhydrous l-galactose
(400 mg, 2.22 mmol) were suspended in dry acetone (9mL)
and treated with concentrated H2SO4 (40 mL). The result-
ing mixture was shaken at room temperature for 24 h. The
cupric sulfate was removed by ®ltration and washed with
acetone. The combined organic phases were neutralized by
addition of Ca(OH)2, ®ltered and concentrated. Puri®ca-
tion by column chromatography (EtOAc:hexane, 1:1) gave
the desired product as a colorless oil (530 mg, 91%). Rf 0.42
(EtOAc:hexane, 2:1). 1H NMR (400 MHz): d=1.32 (3H, s,
CH3), 1.33 (3H, s, CR3), 1.39 (3H, s, CH3), 1.51 (3H, s,
CH3), 3.62 (1H, dd, J=7.3, 11.3 Hz, Ha-6), 3.67 (1H, dd,
J=5.4, 11.3 Hz, Hb-6), 3.83±3.88 (1H, m, H-5), 4.28 (1H,
dd, J=1.9, 8.1 Hz, H-4), 4.34 (1H, dd, J=2.4, 5.1Hz, H-2),
4.62 (1H, dd, J=2.4, 7.8 Hz, H-3), 5.49 (1H, d, J=5.1 Hz,
H-1). 13C NMR (100MHz): d=21.9, 22.6, 23.7, 23.8
(CH3), 59.5, 2Â67.3, 69.4, 69.6, 95.1 (C-1), 107.1, 107.7
(Cq). HRMS (FAB): calcd for C12H20O6 [(M+Na)+]
283.1158, found 283.1155.
Inhibition studies of FucT-III, FucT-V, FucT-VI,
FucT-VII, a-1,3-GalT, b-1,4-GalT, and 2,6ST with four
¯uorinated donor analogues are described.
Materials and Methods
General methods
NMR spectra were recorded on Bruker AM-250, AMX-
400 or AMX-500 spectrometers. 1H chemical shifts are
referenced to residual protic solvent (CDCl3, dH=7.26; -
DMSO-d6, dH=2.50; D2O dH=4.80) or internal standard

M. D. Burkart et al. / Bioorg. Med. Chem. 8 (2000) 1937±1946
1939
Scheme 1. Synthesis of UDP-6F-l-Gal (1): (a) CuSO4, H2SO4 (cat), acetone, 91%; (b) DAST, collidine, DCM, 77%; (c) AcOH, H2O, 68%; (d) BzCl,
pyridine, 70%; 9e) HBr/HOAc, DCM, Ac2O; (f) Ag2CO3, (BnO)2P(O)OH, 3 A ms, DCM, Et2O, CH3CN, 75% over two steps; (g) Pd/C, H2 toluene,
pyridine, NEt3, 95%; (h) MeOH, cyclohexylamine, 87%; (i) BioRad AG 50 W-X2 (NEt3-form), quant, then guanosine morpholidate, tetrazole, pyridine, 2
days, 67%.
(B) 1,2,3, 4-Di-O-isopropylidene-6-deoxy-6-¯uoro-L-galac-
topyranoside (3). A solution of 1,2,3,4-Di-O-isopropyli-
dene-l-galactopyranoside (460 mg, 1.75 mmol) in DCM
(8 mL) was treated with collidine (463 mL, 3.51 mmol)
and DAST (463 mL, 3.51 mmol) at room temperature
and stirred at re¯ux for 6 h. It was diluted with Et2O
and washed with NaHCO3 (2Â), 0.5 N HC1, and brine,
dried over MgSO4 and concentrated to give the ®nal
product 3 as a colorless oil (360 mg, 77%). Rf 0.72
(EtOAc:hexane, 2:3). 1H NMR (400 MHz): d=1.32 (6H,
s, CH3), 1.43 (3H, s, CH3), 1.53 (3H, s, CH3), 4.02±4.10
(1H, m, H-5), 4.25 (1H, dd, J=2.2, 8.1 Hz), 4.33 (1H,
dd, J=2.7, 5.2 Hz), 4.44 (1H, ddd, J=7.0, 9.7, 47.8 Hz,
Ha-6), 4.58 (1H, ddd, J=5.2, 9.5, 46.2 Hz, Hb-6), 4.60±
4.63 (1H, m), 5.53 (1H, d, J=5.1 Hz, H-1). 13C NMR
(100MHz): d=24.4, 24.9, 25.9, 26.0 (CH3), 66.6 (d,
J=23.3 Hz, C-5), 70.3, 70.4, 70.5 (C-2, C-3, C-4), 82.0 (d,
J=168.7 Hz, C-6), 96.1 (C-1), 108.8, 109.6 (Cq). 19F
(367MHz): d= 154.9. HRMS (FAB): calcd for C12
H19FO5 [(M+H)+] 263.1295, found 263.1292.
at 0 ꢁC, and it was stirred for 3 h at room temperature.
The mixture was poured onto ice water and extracted
with EtOAc. The extracts were successively washed with
water, NaHCO3, and brine, dried over anhydrous
MgSO4 and concentrated to give the title compound,
which was used without further puri®cation. Ag2CO3
(138 mg, 0.50 mmol) was added in one portion to a
cooled suspension of anomeric bromide, dibenzylpho-
sphate (209 mg, 0.75 mmol) and 3 A ms (450 mg) in
DCM, Et2O, CH3CN (1.5 mL each) in a round bottom
¯ask lapped with aluminum foil to shut out light. The
mixture was stirred for 15 h at room temperature and
®ltered through Celite, and the ®ltrate was con-
centrated. The residue was chromatographed on silica
gel (toluene:EtOAc, 3:1) to give 5 as a colorless oil
(145mg, 75% over two steps). Rf 0.51 (EtOAc:toluene,
1
1:2). H NMR (400MHz): d=4.34±4.41 (1H, m, H-5),
4.51 (1H, d, J=5.6 Hz), 4.63 (1H, d, J=5.9 Hz), 4.79 (1H,
dd, J=7.0, 11.6 Hz, CHaHbPh), 4.89 (1H, dd, J=6.7,
11.6Hz, CHaHbPh), 5.10 (1H, dd, J=7.8, 11.9 Hz,
CHaHbPh), 5.14 (1H, dd, J=7.3, 11.6 Hz, CHaHbPh),
5.59 (1H, dd, J=3.2, 10.2 Hz), 5.72 (1H, t, J=7.6 Hz, H-1),
5.90±5.96 (2H, m), 7.02±7.05 (2H, m, Harom), 7.18±7.37 (12
H, m, Harom), 7.42±7.53 (4H, m, Harom), 7.61±7.66 (1H, m,
Harom), 7.78±7.82 (2H, m, Harom), 7.94±7.97 (2H, m,
Harom), 8.06±8.09 (2H, m, Harom). 13C NMR (100MHz):
d=67.5 (d, J=5.8 Hz, C-4), 69.5 (d, J=4.4, CH2Ph),
69.6 (d, J=8.7 Hz, C-2), 69.7 (d, J=5.8 Hz, CH2Ph),
71.3 (C-3), 73.1 (d, J=23.3 Hz, C-5), 80.6 (d,
J=173.1 Hz, C-6), 97.0 (d, J=4.4, Hz, C-1), 127.5, 127.9,
128.3, 128.4, 128.5, 128.7, 129.7, 129.8, 130.0 (signal
overlap, Carom), 133.4, 133.6, 133.8 (ipso-Carom), 165.2,
165.4 (signal overlap, CˆO). 19F (376 MHz): d= 155.7.
31P (162 MHz): d= 2.24. HRMS (FAB): calcd for
C41H36FO11P [(M+Cs)ꢂ] 887.1034, found 887.1057.
(C) 1, 2,3,4-Tetra-O-benzoyl-6-deoxy-6-¯uoro-L-galacto-
pyranoside (4). A solution of 6-deoxy-6-¯uoro-l-galac-
tose (150 mg, 0.823 mmol) in dry pyridine (3 mL) was
treated with BzCI (478 mL, 4.12 mmol) and stirred under
argon at room temperature for 16 h. It was diluted with
EtOAc, washed with 0.5 N HCl, NaHCO3, H2O, and
brine, dried (MgSO4) and concentrated. Puri®cation by
column chromatography (EtOAc:hexane, 1:3) gave 4 as
a colorless foam (410 mg, 83%). Rf 0.54 (EtOAc:hexane,
1
1:2). H NMR (400 MHz): d=4.50±4.55 (1H, m, H-6),
4.62±4.67 (1H, m, H-6), 4.74±4.80 (1H, m, H-5), 6.03
(1H, dd, J=3.5, 10.8 Hz), 6.10 (1H, dd, J=3.2, 10.5 Hz),
6.12±6.15 (1H, m), 6.96 (1H, d, J=3.5 Hz, H-1), 7.25±
7.33 (4H, m, Harom), 7.42±7.56 (6H, m, Harom), 7.62±7.68
(2H, m, Harom), 7.81±7.88 (4H, m, Harom), 8.09±8.16 (4H,
m, Harom). 13C NMR (100 MHz): d=67.6 (C-2, C-3),
63.3 (d, J=4.4 Hz, C-4), 68.5 (C-2, C-3), 70.1 (d,
J=23.3 Hz, C-5), 80.8 (d, J=171.6 Hz, C-6), 90.5 (C-1),
128.3, 128.4, 2Â128.5, 129.7, 130.0 (signal overlap, Carom),
133.4, 133.5, 133.8, 134.0 (ipso-Carom), 164.5, 165.4, 165.7,
165.7 (CˆO). HRMS (FAB): calcd for C34H27FO9
[(M+Na)+] 621.1537, found 621.1557.
(E) Di(trimethylammonium)phosphoryl-2,3,4-tri-O-benz-
oyl-6-deoxy-6-¯uoro-ꢀ-L-fucopyranoside (6). A solution
of 5 (300 mg, 0.398 mmol) in toluene (2.4 mL) was trea-
ted with triethylamine (360 mL) and pyridine (480 mL).
The mixture was hydrogenated in an H2 atmosphere
with 10% activated Pd/C (40 mg) as catalyst. After 24h
the mixture was ®ltered and the ®ltrate concentrated to
give 6 as a colorless foam (290mg, 95%). Rf 0.44
(iProp:NH3 aq:H2O, 6:3:1). 1H NMR (500MHz): d=1.21
(3H, t, J=7.4 Hz, N(CH2CH3)3), 3.00 (2H, q, J=7.4Hz,
N(CH2CH3)3), 4.43±4.50 (1H, m, H-5), 4.61 (1H, ddd,
J=5.5, 9.5, 46.4 Hz, Ha-6), 4.65 (1H, ddd, J=5.9, 9.5,
(D) Dibenzylphosphoryl 2,3,4-tri-O-benzoyl-6-deoxy-6-
¯uoro-ꢀ-L-galacto-pyranoside (5). To a cooled solution
of 4 (150 mg, 0.25 mmol) in DCM (1.5 mL) and Ac2O
(150 mL) was added dropwise 30% HBr-ACOH (600 mL)

1940
M. D. Burkart et al. / Bioorg. Med. Chem. 8 (2000) 1937±1946
46.8 Hz, Hb-6), 5.60 (1H, t, J=7.8 Hz, H-1), 5.72 (1H, dd,
J=3.4, 10.3Hz, H-3), 5.76 (1H, dd, J=7.8, 10.3Hz, H-2),
5.93 (1H, bd, J=3.4 Hz, H-4), 7.20±7.25 (2H, m, Harom),
7.35±7.55 (6H, m, Harom), 7.63±7.67 (1H, m, Harom), 7.69±
7.72 (2H, m, Harom), 7.89±8.05 (4H, m, Harom). 13C NMR
(125MHz): d=14.6, 36.1 (N(CH2CH3)3), 60.8 (d,
J=4.8Hz, C-2), 63.2 (d, J=9.5 Hz, C-4), 63.7 (d,
J=22.9 Hz, C-5), 64.5 (C-3), 73.2 (d, J=169.9 Hz, C-6),
88.1 (d, J=4.8 Hz, C-1), 120.1, 120.2, 120.6, 121.3,
121.5, 121.2 (signal overlap, Carom), 125.2, 125.3, 125.6
(ipso-Carom), 157.4, 157.8, 158.1 (CˆO). 19F (376MHz):
d= 40.4. 31P (162 MHz): d=1.20. HRMS (FAB): calcd
for C27H24FO11P [(M+Na)+] 597.0938, found 597.0915.
stirred at room temperature. After 2 days, the mixture
was diluted with water (3 mL) to become a clear solution,
and it was extracted with DCM (3Â5 mL). The aqueous
phase was concentrated and applied to a Bio-Gel P-2
column (2.5Â70 cm), eluted with 50 mM NH4HCO3, to
give the ®nal product 1 as a white solid after lyophili-
zation (53.2 mg, 67%). Rf 0.25 (iProp:1 M NH4OAc,
2:1) 1H NMR (400 MHz): d=3.60 (1H, dd, J=7.6,
10.0 Hz, H-200), 3.68 (1H, dd, J=3.2, 10.0 Hz, H-300),
3.94 (1H, d, J=3.2 Hz, H-400), 3.95±3.99 (1H, m, H-500),
4.18±4.21 (2H, m, H-50), 4.30±4.34 (1H, m, H-40), 4.48±
4.52 (1H, m, H-30), 4.52±4.57 (1H, m, H-600), 4.65±4.69
(1H, m, H-600), 4.73 (1H, t, J=5.7 Hz, H-20), 4.96 (1H, t,
J=7.8 Hz, H-100), 5.88 (1H, d, J=6.0 Hz, H-10), 8.07
(1H, s, H-8). 13C NMR (100 MHz): d=67.8 (d,
J=4.3 Hz, C-50), 70.5 (d, J=5.8 Hz, C-200, C-400), 72.9 (C-
30), 73.6 (d, J=7.2 Hz, C-200, C-400), 74.5 (C-300), 76.1 (d,
J=24.7Hz, C-500), 76.2 (C-200), 85.1 (d, J=165.8 Hz, C-600),
86.2 (d, J=10.1 Hz, C-40), 89.3 (C-10), 100.9 (d, J=7.3 Hz,
C-10), 118.6 (C-5), 140.0 (C-8), 154.2 (C-4), 156.3 (C-2),
161.3 (C-6). 19F (376 MHz): d= 161.0. 31P (162MHz):
d= 12.3 (d, J=19.5 Hz), 10.4 (d, J=20.7 Hz). MS
(ESI): calcd for C16H24O11FN5P2 [(M H) ] 606, found
606; [(M 2H)2 ] 303, found 303.
(F) Di(cyclohexylammonium)phosphoryl 6-¯uoro-ꢀ-L-
fucopyranoside (7). The above phosphate 6 (280 mg,
0.367 mmol) was dissolved in dry MeOH (3 mL) and
cyclohexylamine (3 mL) was added. It was heated at
re¯ux for 4 h. When quantitative debenzoylation was
achieved the mixture was concentrated and fractioned
between water and CHCl3. The water phase was washed
three times with CHCl3 and then concentrated to dryness
by coevaporating MeOH. The product was dissolved in
hot EtOH and precipitated by addition of acetone. Fil-
tration gave 7 as a crystalline colorless solid (147 mg,
87%). Rf 0.36 (iProp:NH3 aq:H2O, 6:3:1). 1H NMR
(400 MHz): d=1.08±1.14 (2H, m, C6H11NH2), 1.23±1.34
(8H, m, C6H11NH2), 1.56±1.62 (2H, m, C6H11NH2),
1.71±1.78 (4H, m, C6H11NH2), 1.91±1.98 (4H, m,
C6H11NH2), 3.05±3.11 (2H, m, C6H11NH2), 3.51 (1H,
dd, J=7.8, 10.0 Hz, H-2), 3.65 (1H, dd, J=3.3, 9.9 Hz,
H-3), 3.91 (1H, d, J=2.4 Hz, H-4), 3.93±3.96 (1H, m,
H-5), 4.52±4.58 (1H, m), 4.63±4.67 (1H, m), 4.83 (1H, t,
J=7.7Hz, H-1). 13C NMR (100MHz): d=26.3, 26.8,
32.8, 52.8 (signal overlap, C6H11NH2), 70.6 (d, J=7.2Hz,
C-2, C-4), 74.4 (d, J=5.8 Hz, C-2, C-4), 74.8 (C-3), 75.9
(d, J=20.2 Hz, C-5), 85.5 (d, J=163.2 Hz, C-6), 100.0 (d,
J=2.9Hz, C-1). 31P (162MHz) d=2.60. 19F (3.76 MHz)
d= 162.5. HRMS (FAB): calcd for C6H12FO8P
[(M+2Cs)+] 526.8284, found 526.8264.
Synthesis of guanosine 50-diphospho-2-deoxy-2-¯uoro-ꢁ-
L-fucopyranoside (GDP-ꢁ2F-Fuc) (8) (Scheme 2)
(A) (2,3-dibenzoyl-2-¯uoro-ꢁ-1-fucopynanosyl) dibenzyl
phosphate (10). This was prepared from dibenzoyl fucal
(9) by our published procedure.27
(B) Di(cyclohexylammonium) 2-deoxy-2-¯uoro-a-fuco-
pyranosyl phosphate (12). 10 was fully deprotected fol-
lowing the same procedure as its b analogue.8 1H NMR
(600 MHz, D2O): 5.54 (dd, J1-P=7.4 Hz, J1-2=4.0Hz, 1H,
1); 4.55 (dddd, J1-2=3.5 Hz, J3-2=10.1 Hz, J2-1=2.2 Hz,
J2-F=50.0 Hz, 1H, 2); 4.20 (q, J5-6=6.6 Hz, 1H, 5); 4.09
(ddd, J3-2=10.1 Hz, J3-F=13.1 Hz, J3-1=3.5 Hz, 1H, 3);
3.81 (t, JF-4=J3-1=3.5 Hz, 1H, 4); 1.13 (d, J5-6=6.6 Hz,
3H, 6); 13C NMR (150 MHz, D2O): 92.86 (dd, JC-P
=5.5 Hz JC-F=22.0 Hz, 1); 89.16 (dd, JC-P=3.7Hz, JC-F
=184.12 Hz, 2); 73.15 (d, J=8.2 Hz, 4); 68.65 (d,
J=17.4Hz, 3); 68.23 (5); 16.01 (6). 31P NMR (161 MHz,
D2O): 0.17; 19F NMR (376MHz, D2O): 202.31 (dd,
J=12.7 Hz, J=50.8 Hz).
(G) Guanosine 50-diphospho-6-¯uono-ꢀ-L-fucopyranoside
(1). 7 (60.0 mg, 0.130 mmol) was dissolved in water
(2mL), applied to a Bio-Rad AG 50 W-X2 cation-
exchange column (Et3N+-form, 2.5Â6 cm), and eluted
with water (150mL). The solution was evaporated, coeva-
porated with MeOH (2Â5 mL), and dried for 3 days under
vacuum to give tnethyiammonium b-l-fucopyranosyl
phosphate. After addition of 4-morpholino-N-N0-dicyclo-
hexylcarboxamidinium guanosine 50-monophosphomor-
pholidate (151mg, 0.209mmol) it was coevaporated with
dry pyridine (3Â5 mL). 1H-Tetrazole (29 mg) and dry
pyridine (1.25 mL) were added, and the solution was
(C) Uridine 50-diphospho-2-deoxy-2-¯uoro-a-fucopyranose
(8). Coupling of 12 with GMP-morpholidate and catalytic
1-H-tetrazole carried out via the procedure by Wittmann.28
1H NMR (600 MHz, D2O): 8.07 (s, 1H), 5.84 (d, J1 -2
0
0
=7.4 Hz, 1H, 10), 5.64 (dd, J1-P=7.0 Hz, J1-2=3.5 Hz, 1H,
1); 4.55 (bdddd, J1-2=J2-P=2.2 Hz, J3-2=10.0 Hz, J2-F
Scheme 2. Synthesis of GDP-2-deoxy-2-¯uoro-a-l-fucose (8); (a) F-TEDA-OTf then (BnO)2POOH (54%); (b) H2 Pd/C; (c) cyclohexylamine (92%
over two steps); (d)

M. D. Burkart et al. / Bioorg. Med. Chem. 8 (2000) 1937±1946
1941
=49.5 Hz, 1H, 2); 4.44 (dd, J=3.1 Hz, J=4.8 Hz, 1H);
4.26 (bs, 1H), 4.18 (q, J5-6=6.6 Hz, 1H, 5); 4.12 (bt,
J=4.9 Hz, 1H), 4.07 (ddd, J3-2=10.1 Hz, J3-F=12.7 Hz,
J3-4 =3.5 Hz, 1H, 3); 3.74 (t, JF-4=J3-4=4.0Hz, 1H); 1.07
(d, J5-6=6.6 Hz, 3H, 6); 31P NMR (161 MHz, D2O):
10.85 (d, J=21.4Hz), 12.51 (d, J=21.4 Hz); 19F NMR
(376 MHz, D2O): 202.53 (dd, J=12.7 Hz, J=53.3 Hz).
and stirred at 25 ꢁC for 3 h. To the solution was added
500 mL 1 M Tris±HCl, pH 7.5, and the solution was
equilibrated to pH 7.5 with 10 N NaOH. To this solution
was added 10 mM MgCl2, 5 mM b-mercaptoethanol,
1 mM ATP, 64 mM acetylphosphate, and 1 mg/mL
bovine serum albumin in a total volume of 5 mL at pH
7.5. 1 Unit of galactokinase and 50 units of AK were
added, and solution was incubated at 25 ꢁC. Each 24 h,
addition of 100 mL 1 M AcOP, re-equilibration to pH
7.5, and addition of 1 U GK and 50 U AK were repe-
ated for 4 days. The solution was adjusted to pH 8.5
with 10 N NaOH. 4.1 mM UTP, 0.25mM UDP-glucose,
and 2.5 mM b-mercaptoethanol were added to the solu-
tion with a total volume of 10mL with pH 8.5. 10 U
inorganic pyrophosphatase, 5 U UDP-glucose pyr-
ophosphory-lase, and 5 U galactose-1-phosphate are
then added. The reaction was allowed to incubate at
25 ꢁC for 3 days with addition of 400 mL UTP (followed by
equilibration to pH 8.0), 5 U UDP-glucose pyropho-
sphorylase, and 5 U galactose-1-phosphate uridyl-
transferase each 24h. The reaction was halted by addition
of 50 U alkaline phosphatase and incubated at room
temperature overnight. The product was puri®ed by ion-
exchange chromatography (Dowex 1-X8, Cl form,
2Â10 cm) with a linear gradient of LiCl (0!0.4 M) in
3 mM HCl. Size-exclusion chromatography (BioRad P-
2, H2O, 2Â30 cm) was used to desalt the product and
gave pure 13 in 60% overall yield. 1H NMR (D2O,
500 MHz) d 7.903 (d, H600, J=8.5 Hz, 1H), 5.935 (d, H10,
J=4.0Hz, 1H), 5.909 (d, H500, J=8.5 Hz, 1H), 5.757 (dd,
H1, J=7.5, 4.0 Hz, 1H), 4.633 (dddd, H2, J=49.5, 10.0,
3.5, 3.0 Hz, 1H), 4.313 (m, H20 and H30, 2H), 4.223 (m,
H40, 1H), 4.154 (m, H3, H5 and H50, 4H), 4.030 (m, H4,
1H), 3.698 (ABq, H6a, J=12.0, 7.0 Hz, 1H), 3.657 (ABq,
H6b, J=12.0, 5.0Hz, 1H). HRMS (FAB) calcd for C15H21
N2O16FP2Na2 (M+H+) 613.0224. found 613.0229.
Two-step-synthesis of uridine 50-diphospho-2-deoxy-2-
¯uoro-ꢁ-D-galactose (UDP-2F-Gal) (13) (Scheme 3).
(A) 2-Deoxy-2-¯uono-ꢁ-D-galactopyranosyl phosphate
(15). Galactal (14) (100 mg, 684 mmol) and Select¯uor
tri¯ate (346 mg, 719 mmol) are dissolved in 3 mL water
and stirred 3 h at 25 ꢁC. To the solution is added 1 mL 1M
HEPES bu€er, pH 7.4, and the solution is adjusted to pH
7.4 with 10 N NaOH and diluted to 10mL. 10mM
MgCl2, 5 mM b-mercaptoethanol, 1 mg/mL bovine albu-
min, 1 mM ATP, 50 mM phosphoacetate, 1 U galacto-
kinase, and 50 U acetate kinase are added, and the
solution is rocked for at 25 ꢁC with additional aliquot of
phosphoacetate added after 48 h. The reaction is mon-
itored by TLC (isopropanol:H2O:30% NH4OH, 7/2/1).
.
After 4 days, BaCl2 H2O (254.8 mg, 1.04 mmol) was
added and the solution was stirred at 4 ꢁC for 6 h. The
obtained white precipitate was removed by centrifugation
and the precipitate was further washed with H2O
(6 mL). After the supernatant was collected, acetone (1
vol) was added and the cloudy solution was allowed to
stand for a day at 4 ꢁC. The solution was centrifuged
and the collected precipitate was washed with cold H2O:
acetone (1:1, 5 mLÂ2) and acetone (5 mL). Treatment of
the resulting barium salt of 14 with Dowex 50 Na+
form followed by drying in vacuo yields the sodium salt
of 14 (131 mg, 63%). 1H NMR (D2O, 500 MHz) d 5.613
(dd, H1, J=8.3, 3.8 Hz, 1H), 4.596 (dddd, H2, J=50.0,
10.0, 3.8, 1.8 Hz, 1H), 4.118 (m, H3 and H5, 2H), 4.007
(m, H4, 1H), 3.697 (m, H6, 2H).
Radiolabeled ꢁ-1,3-fucosyltransferase activity assay.
The activity of FucT V was detected by the GDP-
[U-14C]-fucose assay described previously in which GDP-
[U-14C]-fucose is separated from [U-14C]-fucosylated
product by anion exchange chromatography. Typically,
assays contained 10 mM MnCl2, 0.3 m unit of puri®ed
FucT V and 25 mM cacodylate bu€er (pH 6.2) in a total
assay volume of 0.05 mL. Assays were performed at
25 ꢁC. Reactions were halted with the addition of 0.5 mL
of distilled, de-ionized water. GDP-Fuc was separated
from the product, Lewis x, with a 1.0 mL Dowexl-X8
pipette column. The reaction mixtures were applied to
the column and washed with 0.3 mL distilled, de-ionized
water three times. The ¯ow through and the column
washes were collected in 10mL of ScintiVerse I scintilla-
tion cocktail. Control reactions, without enzyme, were
used to establish the background, non-enzymatic cleavage
rate. A typical control reaction of 32672 cpm guanosine-
50-diphospho-l-[U-14C]-fucose would result in 173 cpm of
non-enzymatic column ¯ow through. GDP-Fuc con-
centration was varied (10, 25, 50, 100 mM) at ®xed con-
centrations of GDP-2F-Fuc (0, 20, 40, 80 mM), and the
acceptor sugar, LacNAc-b-O-(CH2)5CO2CH3, was kept at
twice its Km level, 0.6 mM. Each assay contained 0.3 m unit
of FucT V and 10mM MnCl2 in a 100 mM MES bu€er
(pH 6.0). Reactions were 30min at room temperature. The
(B) Unidine 50-diphospho-2-deoxy-2-¯uoro-ꢁ-D-galactose
(13). Compound 15 was enzymatically converted to 13
in 60% yield following published procedures.19
One-pot procedure for UDP-2-deoxy-2-¯uoro galactose
(13). Galactal (14) (100 mg, 0.68 mmol) and Select¯uor
tri¯ate (346 mg, 0.72 mmol) were dissolved in 3 mL H2O
Scheme 3. One-pot chemo-enzymatic synthesis of UDP-2-deoxy-2-
.
¯uoro-galactose (13): (a) F-TEDA-CH2Cl 2Otf, 3 mL H2O, 3 h;
100 mM Tris±HCl pH 7.4, adjust to pH 7.4 w/10 N NaOH; galacto-
kinase, acetate kinase, ATP, AcOP, bME, BSA, MgCl2, 4 days. 65%:
(b) Galactose-1-phosphate uridyltransferase, UDP-glucose pyrophos-
phorylase, pyrophosphatase, UTP, pH 8.4, 3 days. 67%: (c) F-TEDA-
.
CH2Cl 2Otf, 3 mL H2O, 3 h; 100 mM Tris±HCl pH 7.4, adjust to pH 7.4
w/10 N NaOH; galactokinase, acetate kinase, ATP, AcOP, BME, BSA,
MgCl2, 4 days. Galactose-1-phosphate uridyltransferase, UDP-glucose,
pyrophosphorylase, pyrophosphatase, UTP, pH 8.4, 3 days, 60%.

1942
M. D. Burkart et al. / Bioorg. Med. Chem. 8 (2000) 1937±1946
precise Ki was determined with a nonlinear least square
®t of the data to the equation for competitive inhibition.
was deprotected, benzoylated, and brominated at the
anomeric position. Nucleophilic addition of dibenzyl
phosphate with catalytic Ag2CO3 yields the protected
phosphate 5 in 52% yield over four steps. Complete
deprotection and ion exchange is followed by GMP-
morpholidate using catalytic 1-H-tetrazole,28 to give 1
with a 55% yield over three steps.
Inhibition of ꢁ1,3-GaIT and ꢀ,4-GalT by UDP-2F-Gal.
Assays contained 10mM MnCl2, 0.034 m unit of galacto-
syltransferase, UDP-[6-3H]-galactose (1 to 100 mM),
GlcNAcbOMe (3 mM), inhibitor (1 to 100 mM), bovine
serum albumin (1mg/mL), and 50mM HEPES bu€er
(pH 7.4) in a total assay volume of 0.05 mL. Assays
were performed at 25 ꢁC. Reactions were halted by the
addition of 0.3 mL of deionized water and immediately
applying it to a pipette column of Dowex 1Â8 (1 mL).
The pipette was washed with an additional 0.9 mL of
water (3Â0.3 mL). The ¯ow-through and column
washes were collected in 11 mL of ScintiVerse I scintil-
lation cocktail. A control reaction without enzyme was
used to establish the background count. The precise Ki
was determined with a nonlinear, least square ®t of the
data to the equation for competitive inhibition
Synthesis of GDP-ꢁ-2F-fucose. To understand the ste-
reochemical requirements of the fucosyltransferase
enzymes, we chose to synthesize the opposite anomeric
stereoisomer of GDP-2F-fucose and investigate its inhibi-
tory properties. An equal or greater inhibition of this
molecule to its b analogue will demonstrate a ¯exible
binding domain. The synthesis of this molecule was car-
ried out using a new ¯uorination±phosphorylation tech-
nique from the protected fucal with Select¯uor tri¯ate.27
Following deprotection to 2-deoxy-2-¯uoro fucose-1-
phosphate, coupling with GMP morpholidate was con-
ducted according to published coupling procedures.28
Inactivation studies of ꢁ1,3-GalT. Assays run as above,
with the enzyme samples preincubated as concentrated
solutions (50Â) in di€erent concentrations of UDP-2F-
Gal (0, 1, 0.1, 0.01 mM). Reaction initiated by addition
of enzyme after di€ering times of preincubation (0.5, 4,
10, and 20 mm).
Dibenzoyl fucal (9) was reacted with Select¯uor tri¯ate
and dibenzyl phosphate to yield the ¯uorophosphate 10
in 54% yield. Full deprotection and ion exchange (92%)
was followed by GMP-morpholidate coupling with 1-H-
tetrazole28 to give 8 in a 56% yield.
Inhibition of ꢁ2,6-sialyl transferase by CMP-3F-Neu5Ac.
The assay mixture consisted of 50 mM CMP-Neu5Ac
with 250,000 cpm of CMP-[14C]Neu5Ac added as tracer,
10 mM MnCl2, 0.1% Triton CF-54, 20 mM cacodylate,
pH 6.0, and 0.5 mU of a 02,6-SialylT, and inhibitor (1 to
100 mM) in a 50-mL reaction mixture. Assays were per-
formed at 37 ꢁC for 30 min. Reactions were halted by the
addition of 0.5 mL of 100 mM phosphate bu€er, pH 7.0
and immediately applied to a pipette column of Dowex
1Â8 (1 mL). The pipette was washed with an additional
1 mL of phosphate bu€er (2Â0.5 mL). The ¯ow-through
and column washes were collected in 15 mL of Scinti-
Verse I scintillation cocktail. A control reaction without
enzyme was used to establish background count. The
precise Ki was determined with a nonlinear, least square
®t of the data to the equation for competitive inhibition.
Improved synthesis of UDP-2F-galactose. We sought to
improve the chemoenzymatic synthesis of 13 to facilitate
the large-scale synthesis of this galactosyltransferase
inhibitor. Speci®cally, we wished to develop a one-pot
procedure in which ¯uorination and phosphorylation
and nucleotide transfer could be carried out without
intermediate puri®cation. Recently, Gastinel et al. solved
and re®ned the 3D structure of the catalytic domain of
bovine b-1,4-GalT to 2.4 A.30 Co-crystallization and
resolution of 13 bound to the enzyme active site will
help de®ne the amino acids involved in catalysis.
Although high-yielding enzymatic syntheses of the natural
substrate UDP-Gal have been published,31 a one-pot
procedure has not been reported.
Previous syntheses of 2-deoxy-2-¯uoro galactose have
relied upon reductive ¯uorination of galactal with XeF2-
BF3 OEt2,19 displacement ¯uorination of the selectively-
.
protected talopyranoside with DAST,32 or electrophilic
¯uorination of the tri-O-acetyl galactal with AcOF.33
With the recent development of a one-pot, ¯uorination
technique using the electrophilic ¯uorinating reagent
F-TEDA-CH2Cl (Select¯uor), the ability to ¯uorinate
unprotected sugars with high yields in water became
available.34 With the absence of undesirable side products,
the result of the reaction with Select¯uor and unprotected
galactal is only TEDA-CH2Cl salt and ¯uorinated
galactose. We reasoned that after the addition of bu€er
and equilibration of pH, enzymatic reactions should
proceed normally.
Results and Discussion
Synthesis of GDP-6F-fucose
We chose to synthesize 1 as a comparison with the pre-
viously described GDP-2-deoxy-2-¯uoro-b-l-fucose (16)
for inhibitory activity against fucosyltransferases. This
experiment should determine a correlation between ¯uor-
ine position and inhibition. The synthesis of this molecule
was accomplished by modi®cation of l-galactose (Scheme
1). The previous synthesis of 168 was used as a model
for this synthesis following synthesis of the ¯uorinated
monosaccharide.
To verify this hypothesis, we employed a coupled enzy-
matic assay for galactokinase activity to detect phosphor-
ylation of 2-deoxy-2-¯uoro galactose. Galactokinase
catalyzes the transfer of phosphorous from ATP to galac-
tose, or 2-deoxy-2-¯uorogalactose in this case, to form
l-Galactose was ®rst protected at the 1,2- and 3,4- posi-
tions with isopropylidene, leaving the 6-OH unprotected.
DAST accomplished selective ¯uorination,29 displacing
the 6-OH to give 3 in a 77% yield. The ¯uoroglycoside

M. D. Burkart et al. / Bioorg. Med. Chem. 8 (2000) 1937±1946
1943
galactose phosphate and ADP.35 The assay monitors
the formation of ADP via pyruvate kinase, lactate
dehydrogenase, phosphoenolpyruvate, and NADH.
Disappearance of NADH is followed by UV spectro-
scopy at 340 nM. After reacting 100 mg gaiactal and 1.1
equivalents of Select¯uor (both BF4 and OTf forms) in
10 mL H2O for 3 h followed by addition of bu€er
(100 mM Tris±HCl) and equilibration of the pH to 7.0,
galactokinase demonstrated activity with the ¯uorinated
galactose from the Select¯uor tri¯ate reaction. The
Select¯uor±BF4 reaction product showed no activity.
We concluded that me TEDA-CH2CI tri¯ate salt is
amenable to bu€ered enzymatic conditions at physi-
ological pH.
(BioRad P-2) was used to desalt the product and gave
pure 13 in 60% overall yield.
Fucosyltransferase inhibition. Fucosyltransferases (FucTs)
biosynthesize a variety of oligosaccharides including the
sialyl Lewis x(SLeX), Lewis x (Lex), Lewis a (Lea), and
sialyl Lewis a (SLea) that are central to numerous cell±
cell interactions.36 39 Five human fucosyltransferases
(FucTIII-FucTVII) have been cloned, two of which are
currently commercially available (FucTV and FucTVI).3
Mammalian FucTs transfer a fucose moiety from the
donor guanosine 50-diphospho-b-d-fucose (GDP-Fuc)
to a carbohydrate acceptor in enzyme-dependent a-1,2-,
a-1,3-, a-1,4-, and a1,6-linkages with inversion of
anomeric con®guration. Several of these enzymes, both
bacterial and mammalian, have been studied in great detail
mechanistically,8,21,26,40,41 and attention has recently been
focused upon the identi®cation of peptide sequence
homology, acceptor binding domains, and catalytic
residues.42 45
Given this positive result, we attempted the one-pot
chemo-enzymatic synthesis of 2-deoxy-2-¯uorogalactose
phosphate (15) (Scheme 1). Galactal (14) and Select¯uor
tri¯ate were dissolved in 3 mL H2O and stirred at 25 ꢁC
for 3 h. To the solution was added 1 M Tri±HCl, pH
7.5, and the solution was equilibrated to pH=7.5 with 10
N NaOH. To this solution was added MgCl2, b-
mercaptoethanol, ATP, phosphoacetate, bovine serum
albumin, galactokinase, and alkaline phosphatase to a
total volume of 5 mL. The solution was incubated at
25 ꢁC for 4 days with addition of 1 equiv phosphoace-
For our experiments, we focused on mammalian fuco-
syltransferases responsible for the biosynthesis of sialyl
trimeric Lewisx, which is a portion of the sugar chains
attached to the P-selectin glycoprotein ligand-1 (PSGL-
1) glycan.46 These include FucTIII, V, VI, and VII.
FucTV and FucTVI are commercially available (Cal-
biochem, San Diego, CA), and FucTIII and FucTVII
were expressed as baculovirus constructs in cultured
insect cells.47 Inhibition assays were performed and the
results are summarized in Table 1. Both 1 and 16 display
competitive inhibtion against all four mammalian fuco-
syltransferases with KIs at or below Km values.44,48,49
Given earlier studies of FucTV, correlating inhibition of
16 with a cationic transition state species,8 and 90%
homology between the studied fucosyltransferases, we
deduced that FucTIII, V, VI, and VI perhaps share
considerable mechanistic features, including formation
of an oxocarbenium-ion species in the transition state
(Fig. 1), but the proposed mechanism remains to be
proven.
.
tate each 24h. BaCl2 2H2O was added to precipitate bar-
ium phosphate. The supernatant was diluted with cold
acetone and incubated at 4 ꢁC to precipitate the product.
The precipitate was collected by centrifugation, washed
with ice cold H2O/acetone and then cold acetone. The
solid was dried in vacuo, and the product was incubated
with H2O and Dowex 50 ion-exchange resin (Na+
form) to give the sodium salt. The ®ltrate was dried in
vacuo to yield 15 in 65% yield.
Finally, a one-pot method was applied to the complete
synthesis of 13. The above procedure to make 15 was
followed without barium precipitation. Instead, the pH
was adjusted to 8.5 with 10 N NaOH. UTP, inorganic
pyrophosphorylase, UDP-glucose pyrophosphorylase,
galactose-1-phosphate uridyltransferase, and a catalytic
amount of UDP-glucose were added to a total volume
of 10 mL. The reaction was allowed to incubate at 25 ꢁC
for 3 days with addition of extra UTP, UDP-glucose
pyrophosphorylase, and galactose-1-phosphate undyl-
transferase after 24h. The reaction was halted by addition
of alkaline phosphatase. The product was puri®ed by ion-
exchange chromatography (Dowex 1Â8, Cl form) with
a linear gradient of LiCl. Size-exclusion chromatography
An interesting issue is whether these competitive inhibi-
tors are non-reacting analogues or slow substrates to the
FucTs. Whereas it is known that GDP-2F-Fuc acts as a
slow substrate for FucTIII, the same molecule is not
accepted as a substrate for FucTV or FucTVI.8,50,51
Within this inhibition study the e€ect of very slow reac-
tivity is negligible within the assay timeframe, the di€er-
ential reactivity between the enzymes is indicative of
Table 1. Inhibition of four human fucosyltransferases by GDP-2F-Fuc, GDP-6F-Fuc, and GDP-a-2F-Fuc
GDP-Fuc Km (mM)
(16) Ki (mM)
(1) Ki (mM)
(8) Ki (mM)
Fuct-III
Fuct-V
Fuct-VI
Fuct-VII
33.6
18.8
9.0
38Æ4
4.0Æ0.6
10Æ2.4
21Æ2
22Æ10
3.4Æ1.0
1.0Æ0.5
11Æ2
Ð
36Æ4
2Æ1
Ð
8.0

1944
M. D. Burkart et al. / Bioorg. Med. Chem. 8 (2000) 1937±1946
Figure 1. Putative transition-state structure of fucosyltransferases.
Figure 2. Putative transition-state structure of sialyltransferase.
possible mechanistic di€erences within the FucT family
and merits further study.
Also of interest is the similarity between the KI values of
the two inhibitors, demonstrating that decoration of the
glycon at the 2- and 6-positions with ¯uorine, equidistant
from the ring oxygen, provide similar electronic e€ect. It
is therefore valuable to note the similar activity of 2-
and 6-¯uoro glycosides for future inhibitor design.
Galactosyltransferase inhibition. b-1,4-Galactosyltrans-
ferase (b-1,4-GalT) has been extensively studied due to
its role in the biosynthesis of blood group antigens and
sialyl Lewis x.19,20,63 65 Changes in the in vivo activity
of b-1,4-GalT have been shown to be indicative of
arthritis and cancer progression.66,67 A role of b-1,4-
GalT in mammalian fertilization as a cell-surface recep-
tor has also been discovered.68,69 a-1,3-GalT, an enzyme
found in pigs and other mammals, is involved in the
biosynthesis of xenoantigens responsible for the rejection
of xenotransplanted organs in humans.70 An important
distinction of a-1,3-GalT is that a `retaining' glycosyl-
transferase, where the anomeric con®guration of the
transferred galactose moiety remains in alpha con®g-
uration, like that of the sugar nucleotide donor uridine
50-diphosphogalactose (UDP-Gal). b-1,4-GalT is an
inverting glycosy1transferase, where the anomeric con-
®guration is reversed in the reaction from a in UDP-Gal
to b in the glycoside product. The `retaining' nature of
a-1,3-GalT implies a two-step process for the transfer
of UDP-Gal to the acceptor as to that of retaining
glycosidases.9
Inhibition of 8 with FucTV and FucTVI showed that
the unnatural anomeric con®guration in the ¯uorinated
analogue provides signi®cant inhibition. This result
indicates that the binding domain of fucosyltransferase
enzymes for the sugar moiety allows stereochemical
¯exibility.
Sialyltransferase inhibition. Sialyltransferases are the
enzymes responsible for transfer of sialic acid from cyti-
dine 50-monophospho-N-acetylneuraminic acid (CMP-
Neu5Ac) to a growing oligosaccharide, often terminat-
ing the series at the reducing end.52 Recent synthesis of
potent inhibitors of a-2,6-sialyltransferase (2,6ST) have
elucidated important conformational structure of the
transition state to include sp2 character at the anomeric
carbon.53 57 However, a cationic transition-state structure
has not yet been demonstrated in mechanistic studies of
these enzymes, although one would be predicted given
the quaternary structure of the anomeric carbon. Only
solvolysis of CMP-Neu5Ac has been shown to display a
®nite sialyl cation species.58 Additionally, N-Acetyl-3-
¯uoro-neuraminic acid alone serves as a competitive
inhibitor of bacterial and viral sialidases.59 61 There-
fore, inhibition of sialyltransferases by a ¯uorinated
CMP-Neu5Ac derivative will provide additional sup-
port for the proposed cationic transition state (Fig. 2).
b-1,4-GalT catalyzes an inversion of con®guration at
the anomeric position of the transferred sugar, from a
con®guration in the sugar nucleotide donor uridine 50-
diphosphogalactose (UDP-Gal) to the transferred b
saccharide. Srivastava et al. showed that the 2-deoxy-
genated derivative of UDP-Gal had kinetic parameters
almost identical to the natural substrate with
b(1,4)GalT (Km=51 mM).71 Given that this hydroxy
group may be necessary for substrate binding, an equal
or better substrate would a€ect the oxocarbenium for-
mation, if existing, imparted by removal of the electro-
negative oxygen at C-2.72 Takayama et al. demonstrated
the inhibition properties of 13 for a-1,3-GalT (Ki=2 mM)
and b-1,4-GalT (IC50=245mM.8,18 A further piece of
mechanistic study could be found in the structural
determination of galactosyltransferase by single crystal
X-ray di€ractometry with a bound inhibitor. Co-crys-
tallization and resolution of UDP-2-deoxy-2-¯uoro-
galactose with this enzyme will help de®ne the amino
We chose to study 2,6ST initially due to its commercial
availability (Calbiochem). Inhibition assays were per-
formed and are summarized in Table 2. Cytidine 50-
monophospho-N-acetyl-3-¯uoro-neuraminic acid (17)62
displayed competitive inhibition against 2,6ST with a KI of
5.7 mM. Compared to the Km for CMP-Neu5Ac of 15mM,
this result is consistent with a transition-state-structure
containing considerable oxo carbenium character.

M. D. Burkart et al. / Bioorg. Med. Chem. 8 (2000) 1937±1946
Table 2. Inhibition of a2,6ST by CMP-3F-Neu5Ac
1945
CMP-Neu5Ac
Km (mM)
(2) Ki (mM)
5.7Æ2
a2,6ST
15
acids involved in catalysis in order to con®rm the
mechanistic hypothesis.
In assaying both a and b galactosyltransferase, we
hoped to understand inherent mechanistic di€erences
between the two classes of enzyme. Table 3 summarizes
the inhibition resufts. The large disparity in IC50 values
between a and b galactosyltransferase (35-fold di€erence)
may be explained by the mechanistic assumptions made
about these enzymes. We can rationalize inhibition of
b(1,4)GalT in a manner similar to fucosyltransferases
and sialyltransferase above. Formation of the oxocarbe-
nium in displacement of the nucleotide pyrophosphate
and prior to addition of the acceptor suciently may
explain this result (Fig. 3).
Figure 3. Putative transition-state structure of b-galactosyltransferase.
Conclusion
Here we have provided useful synthetic methods for the
preparation of ¯uorinated sugar nucleotides which may
®nd use as probes of glycosyltransferases. This study
demonstrates that the general non-retaining glycosyl-
transferase reactions may proceed through a charged
species associated with the donor glycon within the
transition-state structure, most the oxocarbenium-ion
species. Alternatively, these ¯uoro nucleotides may sim-
ply bind to the active site as non-productive substrates,
and therefore are observed as competitive inhibitors.
This trend is seen with human fucosyltransferase III, V,
VI, and VII, a(2,6)sialyltransferase, and b(1,4)galactosyl-
transferase. Also demonstrated is the non-inactivation of
a(1,3)galactosyltransferase, indicating that retaining
glycosyltransferases may not undergo covalent enzyme-
glycon intermediates similar to retaining glycosidases.
These ¯uoro sugar nucleotides should be useful for the
structural study of glycosyltransferases as they form a
stable tight complex with the enzyme.
a(1,3)GalT, on the other hand, has been predicted to
undergo a two-step, double-displacement mechanism in
which a nucleophile is transiently involved in forming a
glycosyl±enzyme complex.73 This mechanism is analogous
to that of retaining glycosidases, which have been con-
clusively demonstrated to form a covalent enzyme±glycon
intermediate.14 Here the leaving group (UDP) and the
acceptor reside on the same side of the glycon, indicating a
synclinal, two-step mechanism. Ichikawa et al. demon-
strated inhibition by a charged iminocyclitol, implicating
interaction with a carboxylate residue to impart strong
binding.74 Because this proposed mechanism includes
the existence of a covalent intermediate between enzyme
and the glycon, we investigated enzyme inactivation by 13
of a(1,3)-GalT.75 2-Deoxy-2-¯uoro glycosides have been
shown to inactivate glycosidases following formation of
an enzyme±glycon covalent intermediate.14 Our studies,
which analyzed inactivation concentrations of 0, 1, 0.1,
and 0.01 mM, showed no enzyme inactivation by the
¯uorinated nucleotide sugar. Hence, the mechanism
may di€er signi®cantly from the formation of a covalent
intermediate in the two-step reaction of glycosidases.
Acknowledgements
This research was supported by the NIH GM44154 and
the NSF. Arno Du‚es thanks the Ernst Schering
Research Foundation for a fellowship.
References
1. Varki, A. Glycobiology 1997, 3, 97.
2. Field, M. C.; Wainwright, L. J. Glycobiology 1995, 5, 463.
3. Paulson, J. C.; Colley, K. J. J. Biol. Chem. 1989, 264,
17615.
4. Withers, S. G.; Can, J. Chem. 1999, 77, 1.
5. Poulter, C. D., In Biomontiers of Fluorine Chemistry;
Ojima, I., McCarthy, J. R., and Welch, J. T., Eds. 1996, p
158.
Table 3. Inhibition of a(1,3)GalT and b(1,4)GalT by UDP-2F-Gal
6. Chapeau, M.-C.; Frey, P. A. J. Org. Chem. 1994, 59, 6994.
7. Thoden, J. B.; Hegeman, A. D.; Wesenberg, G.; Chapeau, M.
C.; Frey, P. A.; Holden, H. M. Biochemistry 1997, 36, 6294.
8. Kim, S. C.; Singh, A. N.; Raushel, F. M. Arch. Biochem.
Biophys 1988, 267, 54.
GDP-Gal
Km (mM)
Ki (mM)
a(1,3)GalT
b(1,4)GalT
17Æ3
3Æ6
245
2.0Æ0.3
9. Murray, B. W.; Wittmann, V.; Burkart, M. D.; Hung, S.-C.;
Wong, C.-H. Biochemistry 1997, 36, 823.

1946
M. D. Burkart et al. / Bioorg. Med. Chem. 8 (2000) 1937±1946
10. Koshland, D. E. Biol. Rev. 1953, 28, 416.
11. Davies, G.; Henrissat, B. Structure 1995, 3, 853.
12. McCarter, J. D.; Withers, S. G. J. Am. Chem. Soc. 1996,
118, 241.
13. Withers, S. G.; Rupitz, K.; Street, I. P. J. Biol. Chem.
1988, 263, 7929.
42. Breton, C.; Oriol, R.; Imberty, A. Glycobiology 1998, 8,
87.
43. Nguyen, A. T.; Holmes, E. H.; Whitaker, J. M.; Ho, S.;
Shetterly, S.; Macher, B. A. J. Biol. Chem. 1998, 273, 25244.
44. Vo, L.; Lee, S.; Marcinko, M. C.; Holmes, E. H.; Macher,
B. A. J. Biol. Chem. 1998, 273, 25250.
14. Withers, S. G.; Street, I. P. J. Am. Chem. Soc. 1988, 110,
8551.
15. Porter, D. J. T.; Merrill, B. M.; Short, S. A. J. Biol. Chem.
1995, 270, 15551.
16. White, A.; Tull, D.; Johns, K.; Withers, S. G.; Rose, D. R.
Nature Structual Biology 1996, 3, 149.
17. Withers, S. G.; Warren, R. A.; Street, I. P.; Rupitz, K.;
Kempton, J. B.; Abersold, R. J. J. Am. Chem. Soc. 1990, 112,
5887.
18. Takayama, S.; Chung, S.-J.; Igarashi, Y.; Ichikawa, Y.;
Sepp, A.; Lechler, R. I.; Wu, J.; Hayashi, T.; Siuzdak, G.;
Wong, C.-H. Bioorg. Med. Chem. 1999, 7, 401.
19. Hayashi, T.; Murray, B.; Wang, R.; Wong, C.-H. Bioorg.
Med. Chem. 1997, 5, 497.
20. Ichikawa, Y.; Lin, Y.-C.; Dumas, D. P. G.-J.S.; Garcia-
Junceda, E.; Williams, M. A.; Bayer, R.; Ketcham, C.;
Walker, L. E.; Paulson, J. C.; Wong, C.-H. J. Am. Chem. Soc.
1992, 114, 9283.
45. Sherwood, A. L.; Nguyen, A. T.; Whitaker, J. M.;
Macher, B. A.; Stroud, M. R.; Holmes, E. H. J. Biol. Chem.
1998, 273, 25256.
46. Wilkins, P. P.; McEver, R. P.; Cummings, R. D. J. Biol.
Chem. 1996, 271, 18732.
47. Shinkai, A. S. K.; Sasaki, K.; Morishita, Y.; Nishi, T.;
Matsuda, Y.; Takahashi, I.; Anazawa, H. Protein Expr. Purif
1997, 10, 379.
48. De Vries, T.; Palcic, M.; Schoenmakers, P.; Van Den
Eijnden, D.; Joziasse, D. Glycobiol. 1997, 7, 921.
49. Britten, C. J.; Bird, M. I. Biochim. Biophys. Acta 1997,
1334, 57.
50. Baisch, G.; Ohrlein, R.; Katopodis, A.; Strei€, M.; Kol-
binger, F. Bioorg. Med. Chem. Lett. 1997, 7, 2447.
51. No conversion was observed at 1 mM concentration in the
presence of 1 mM LacNac.
52. Harduin-Lepers, A.; Recchi, M. A.; Deiannoy, P. Glyco-
biology 1995, 5, 741.
21. Qiao, L.; Murray, B. W.; Shimazaki, M.; Schultz, J.;
Wong, C.-H. J. Am. Chem. Soc. 1996, 118, 7653.
22. Hinsgaul, O.; Kaur, K. J.; Srivastava, G.; Blaszcyk-
Thurin, M.; Crawley, S. C.; Heerze, L. D.; Palcic, M. M. J.
Biol. Chem. 1991, 266, 17858.
23. Lowary, T. L. Carbohydr. Res. 1993, 249, 163±195.
24. Andersen, S. M.; Ebner, M.; Ekhart, C. W.; Gradnig, G.;
Legler, G.; Lundt, I.; Stutz, A. E.; Withers, S. G.; Wrodnigg,
T. Carbohydr. Res. 1997, 301, 155.
25. McCarter, J. D.; Yeung, W.; Chow, J.; Dolphin, D.;
Withers, S. G. J. Am. Chem. Soc. 1997, 119, 5792.
26. Murray, B. W.; Takayama, S.; Schultz, J.; Wong, C.-H.
Biochemistry 1996, 35, 11183.
27. Vincent, S. P.; Burkart, M. D.; Tsai, C.-Y.; Zhang, Z.;
Wong, C.-H. J. Org. Chem. 1999, 64, 5264.
28. Wittmann, V.; Wong, C.-H. J. Org. Chem. 1997, 62, 2144.
29. Picq, D.; Anker, D. Carbohydr. Res. 1987, 166, 309.
30. Gastinel, L.; Cambillau, C.; Bourne, Y. EMBO J. 1999,
18, 3546.
31. Heidlas, J. E.; Lees, W. J.; Whitesides, G. M. J. Org.
Chem. 1992, 57, 152.
32. Haradahira, T.; Maeda, M.; Yano, Y.; Kojima, M. Chem.
Pharm. Bull 1984, 32, 3317.
33. Adam, M. J. J. Chem. Soc., Chem. Commun., 1982, 730.
34. Burkart, M. D.; Zhang, Z.; Hung, S.-C.; Wong, C.-H. J.
Am. Chem. Soc. 1997, 119, 11743.
35. Sabater, B.; Sebastian, J.; Asensio, C. Biochim. Biophys.
Acta 1972, 284, 414.
36. Weston, B. W.; Smith, P. L.; Kelly, R. J.; Lowe, J. B. J.
Biol. Chem. 1992, 267, 24575.
37. Ichikawa, Y.; Halcomb, R. L.; Wong, C.-H. Chem. Br.,
1994, 117±121.
38. Foxall, C.; Watson, S. R.; Dowbenko, D.; Fennie, C.;
Lasky, L. A.; Kiso, M.; Itasagawa, A.; Asa, D.; Brandley, B.
K. J. Cell Biol. 1992, 117, 895.
39. Parekh, R. B.; Edge, C. J. TIBTECH 1994, 12, 339.
40. Hinsdsgaul, O.; Kaur, K. J.; Srivastava, G.; Blaszczyk-
Thurin, M.; Crawley, S. C.; Heerze, L. D.; Palcic, M. M. J.
Biol. Chem. 1991, 266, 17858.
53. Schaub, C.; Muller, B.; Schmidt, R. R. Glycoconjugate J.
1998, 15, 345.
54. Muller, B.; Schaub, C.; Schmidt, R. R. Angew. Chem., Int.
Ed 1998, 37, 2893.
55. Muller, B.; Martin, T. J.; Schaub, C.; Schmidt, R. R.
Tetrahedron Lett. 1998, 39, 509.
56. Kajihara, Y.; Kodama, K.; Wakabayashi, T.; Sato, K.;
Hashimoto, H. Carbohydr. Res. 1993, 247, 179.
57. Kajihara, Y.; Hashimoto, H.; Kodama, H.; Wakabayashi,
T.; Sato, K. J. Carbohydr. Chem. 1993, 12, 991.
58. Horenstein, B. A.; Bruner, M. J. Am. Chem. Soc. 1996,
118, 10371.
59. Hagiwara, T.; Kijima-Suda, I.; Ido, T.; Ohrui, H.; Tomita,
K. Carbohydr. Res. 1994, 263, 167.
60. Petrie, C. R. III; Sharma, M.; Simmons, O. D.; Korytnyk,
W. Carbohyd. Res. 1989, 186, 326.
61. Nakajima, T.; Hori, H.; Ohrui, H.; Meguro, H.; Ido, T.
Agric. Biol. Chem. 1988, 52, 1209.
62. Burkart, M. D.; Vincent, S. P.; Wong, C.-H. Chem. Com-
mun. 1999, 1525.
63. O'Kee€e, E. T.; Hill, R. L.; Bell, J. E. Biochemistry 1980,
19, 4954.
64. Kajihara, Y.; Ebata, T.; Koseki, K.; Kodama, H.; Mat-
sushita, H.; Hashimoto, H. J. Org. Chem. 1995, 60, 5732.
65. Trayer, I. P.; Hill, R. L. J. Biol. Chem. 1971, 246, 6666.
66. Krishnaraj, R.; Saat, Y. A. Cancer Res. 1985, 45, 3615.
67. Alavi, A.; Axford, J.. Glycoconj. J, 206.
68. Gong, X.; Dubois, D. H.; Miller, D. J.; Shur, B. D.
Science 1995, 269, 1718.
69. Barcellow-Ho€, M. G. Exp. Cell Res. 1992, 201, 225.
70. Sharma, A.; Okabe, J.; Birch, P.; McClellan, S.; Martin, M.;
Platt, J.; Logan, J. Proc. Natl. Acad. Sci. USA 1996, 93, 7190.
71. Srivastava, G.; Hindsgaul, O.; Palcic, M. M. Carbohydr.
Res. 1993, 245, 137.
72. Capon, B. Chem. Rev. 1969, 69, 407.
73. Lowary, T. L.; Hindsgaul, O. Carbohydr. Res. 1993, 249,
163.
74. Kim, Y. J.; Ichikawa, M.; Ichikawa, Y. J. Am. Chem. Soc.
1999, 121, 5829.
75. Silverman, R. B. Methods Enzymol. 1995, 249, 241.
41. Palcic, M. M. J. Biol. Chem. 1989, 264, 17174.