Design and synthesis of 2'-deoxy-2'-fluorodisaccharides as mechanism- based glycosidase inhibitors that exploit aglycon specificity

Article abstract of DOI:10.1021/ja9627454

Stable, aglycon-specific inactivators of glycosidases have considerable potential as tools in the study of mechanisms of oligosaccharide processing, and possibly as avenues toward new therapeutics. Glycosidases for which the rate-determining step with the

Full text of DOI:10.1021/ja9627454

5792
J. Am. Chem. Soc. 1997, 119, 5792-5797
Design and Synthesis of 2′-Deoxy-2′-Fluorodisaccharides as
Mechanism-Based Glycosidase Inhibitors That Exploit Aglycon
Specificity
John D. McCarter, Wai Yeung, Jack Chow, David Dolphin, and
Stephen G. Withers*
Contribution from the Department of Chemistry, UniVersity of British Columbia, 2036 Main Mall,
VancouVer, British Columbia, Canada V6T 1Z1
ReceiVed August 7, 1996. ReVised Manuscript ReceiVed April 21, 1997X
Abstract: Stable, aglycon-specific inactivators of glycosidases have considerable potential as tools in the study of
mechanisms of oligosaccharide processing, and possibly as avenues toward new therapeutics. Glycosidases for which
the rate-determining step with the natural substrate is the hydrolysis of the glycosyl-enzyme intermediate are shown
to be inactivated by the 2′-deoxy-2′-fluoro derivative of this substrate. Thus Agrobacterium faecalis â-glucosidase
is inactivated by 2′-deoxy-2′-fluorocellobiose according to inactivation parameters of ki ) 0.018 min-1 and Ki ) 20
mM. Inactivation is shown to occur Via the accumulation of the same 2-deoxy-2-fluoroglycosyl-enzyme intermediate
as that formed using activated 2-deoxy-2-fluoroglycosides by identification of the labeled peptide in proteolytic
digests. Thus, interactions between the enzyme and the sugar aglycon provide sufficient transition state stabilization
to allow formation and trapping of the glycosyl-enzyme. â-Glucocerebrosidase, a â-glucosidase specific for hydrolysis
of glucocerebrosides, is not inactivated by 2′-deoxy-2′-fluorocellobiose, thereby demonstrating the aglycon specificity
of this class of inactivator.
Introduction
aglycon, and essentially full glycon, specificity. Inhibitors of
this class have a single sterically conservative modification,
differing from the natural substrates of the enzyme by the
replacement of a single hydroxyl group by fluorine on an
oligosaccharide which could potentially consist of several sugar
residues of different type and linkage specificity. Such com-
pounds thus have the potential for extremely selective inhibition
of specific glycosidases and glycosyltransferases, which could
be of considerable significance in the design of the therapeutic
agents targeting these enzymes. They should also, unlike
inactivators which include an activated leaving group, be
extremely stable to spontaneous hydrolysis.
Previously reported covalent glycosidase inactivators have
included conduritol epoxides,1 glycosylmethyltriazenes,2 N-
(haloacetyl)glycosylamines,3 and 2-deoxy-2-fluoroglycosides
with good leaving groups such as 2,4-dinitrophenolate or
fluoride.4,5 Each of these inhibitor classes relies upon the
presence of either a reactive functional group or activated
leaving group in the molecule which results in covalent labeling
of the enzyme. Glycon specificity is conferred by structural
similarity of the inactivator to the appropriate sugar in the glycon
portion of the molecule, but none of these inactivators exploit
aglycon specificity, which is sacrificed to permit incorporation
of a reactive or activated, but non-natural, moiety. Glycosidases
featuring multiple subsites, such as the paradigmatic lysozyme,
bind oligosaccharide substrates in two or more adjacent active
site regions. The glycon, or catalytic, subsite contains the
catalytic residues and binds the nonreducing end sugar proximal
to the bond to be cleaved while the aglycon subsite binds the
“aglycon” leaving group portion of the substrate, which may,
depending on the enzyme, be one or more sugars, a complex
oligosaccharide, a fatty acyl group, or various other moieties.
We report the design of a covalent, mechanism-based
inactivator of a retaining glycosidase6 that incorporates full
While most glycosidase inactivators function by the (es-
sentially) irreversible reaction of reactive functional groups with
active site residues, 2-deoxy-2-fluoroglycosides with good
leaving groups inactivate retaining glycosidases by the trapping
of a stabilized 2-deoxy-2-fluoroglycosyl-enzyme rendered ki-
netically accessible by the presence of the activated leaving
group. This leaving group accelerates only the first, glycosyl-
ation step while the C-2 fluorine slows both glycosylation and
deglycosylation steps so that the resulting 2-deoxy-2-fluoro-
glycosyl-enzyme accumulates. With some enzymes, it may be
possible to incorporate the natural aglycon in place of this
unnatural leaving group since the additional binding energy
derived from interactions between the aglycon and the enzyme’s
aglycon-binding subsite may substantially increase the rate of
the glycosylation step, resulting in a natural leaving group of
nucleofugacity effectively comparable to that of unnatural
leaving groups of much lower pKa. The requirement would
therefore be that the deglycosylation step for the natural substrate
be slower than glycosylation, and thus most likely be rate-
limiting. Indeed, with Agrobacterium â-glucosidase, the kcat
value for cellobiose (4-O-(â-D-glucopyranosyl)-D-glucose), the
natural substrate of the enzyme, is 9900 min-1, comparable to
that of p-nitrophenyl â-D-glucoside (10 100 min-1) or 2,4-
dinitrophenyl â-D-glucoside (10 700 min-1),7,8 despite 107-1010-
* To whom correspondence should be addressed. Phone: (604) 822-
3402. FAX: (604) 822-2847. E-mail: withers@chem.ubc.ca.
X Abstract published in AdVance ACS Abstracts, June 1, 1997.
(1) Legler, G. AdV. Carbohydr. Chem. Biochem. 1990, 48, 319.
(2) Sinnott, M. L.; Smith, P. J. J. Chem. Soc., Chem. Commun. 1976,
223.
(3) Naider, F.; Bohak, Z.; Yariv, J. Biochemistry 1972, 11, 3202.
(4) Withers, S. G.; Street, I. P.; Bird, P.; Dolphin, D. H. J. Am. Chem.
Soc. 1987, 109, 7530.
(5) Withers, S. G.; Rupitz, K.; Street, I. P. J. Biol. Chem. 1988, 263,
7929.
(6) Retaining glycosidases catalyze the hydrolysis of glycosidic bonds
with net retention of anomeric configuration, presumably by a double-
displacement mechanism involving attack of an enzymic nucleophile, leading
to formation of a glycosyl-enzyme intermediate (“glycosylation”) and
subsequent hydrolysis of the intermediate (“deglycosylation”).18
S0002-7863(96)02745-X CCC: $14.00 © 1997 American Chemical Society

Design and Synthesis of 2′-Deoxy-2′-fluorodisaccharides
J. Am. Chem. Soc., Vol. 119, No. 25, 1997 5793
Scheme 1
fold differences in leaving group Ka (the pKa of the 4-position
hydroxyl of glucose is ∼14, Versus pKa values of 7.1 and 3.9,
respectively, for p-nitrophenolate and 2,4-dinitrophenolate).
Deglycosylation has been shown to be rate-limiting with this
enzyme for these two aryl glycosides and must therefore also
be rate-limiting for cellobiose.7 Thus binding interactions of
the enzyme with the sugar aglycon increase the effective leaving
group ability of the sugar 107-fold.9 If this binding energy is
sufficient to ensure that the deglycosylation step with a 2′-deoxy-
2′-fluoro disaccharide is rate-limiting, then the resulting inter-
mediate (identical to that found upon release of 2,4-dinitrophe-
nolate or fluoride) would be expected to accumulate, inactivating
the enzyme. In order to test this hypothesis, the fluorinated
disaccharides 4-O-(2-deoxy-2-fluoro-â-D-glucopyranosyl)-D-
glucose (2′-deoxy-2′-fluorocellobiose, 8) and 4-O-(2-deoxy-2-
fluoro-â-D-galactopyranosyl)-D-glucose (2′-deoxy-2′-fluorolac-
tose, 14) were synthesized.
steps (thiolation, benzylation, bromination, and hydrogenolysis
after coupling) but was highly stereoselective in the key
glycosylation step. However, despite the fact that the yield of
the protected â-1,4-linked fluorodisaccharide 19 in the coupling
step was doubled, the stereoselective approach resulted in lower
overall yield (based on the common intermediate 1 of fluoro-
disaccharides due to the extra steps required.
In the low stereoselectivity approach (Scheme 1), the fluo-
roglycosyl bromide 1 was reacted with the anhydro sugar 3 in
the presence of silver carbonate, silver trifluoromethanesulfonate,
and Drierite in acetonitrile. After being stirred overnight at room
temperature, the mixture was filtered and chromatographed on
silica to give a mixture of the previously reported10 R-glycoside
4 and the novel â-1,4-linked fluorodisaccharide 5 along with
unreacted 3 and the hydrolysis product 6. The 2′-deoxy-2′-
fluorodisaccharides were isolated in ∼50% yield (based on the
bromosugar) with an R/â anomeric ratio of ∼70/30 as deter-
mined by 1H-NMR. The protected fluorocellobiose derivative
5 was isolated from the fluorodisaccharide mixture by fractional
crystallization, chromatography, and crystallization. Acetolysis
of the 1,6-anhydro ring of 5 gave a mixture of anomeric acetates
7 which could not be induced to crystallize. Upon careful
chromatography two samples that were enriched in each anomer
were isolated and their 1H-NMR spectra determined. Deacetyl-
ation of 7 by the Zemple´n method11 using sodium methoxide
in methanol gave a quantitative yield of the syrupy 2′-deoxy-
2′-fluorocellobiose (8).
Results and Discussion
Synthesis of â-1,4-Linked 2′-Deoxy-2′-fluorodisaccharides.
2′-Deoxy-2′-fluorocellobiose (8). Two methods were used to
synthesize the title compound 8. The less stereoselective
approach resulted in the desired diastereomeric fluorodisaccha-
ride as the minor diastereomer but required fewer synthetic steps
and gave good mass recovery of fluorodisaccharide anomers.
The stereoselective approach required four additional synthetic
(7) Kempton, J. B.; Withers, S. G. Biochemistry 1992, 31, 9961.
(8) Day, A. G.; Withers, S. G. Can. J. Biochem. 1986, 64, 914.
(9) Based upon the similar enzyme-catalyzed hydrolysis rates observed
for the glycosides of p-nitrophenol (pKa ) 7.1) and glucose (pKa ≈ 14).
(10) Shelling, J. G.; Dolphin, D.; Wirz, P.; Cobbledick, R. E.; Einstein,
F. W. B. Carbohydr. Res. 1984, 132, 241.
(11) Zemple´n, G. Ber. Dtsch. Chem. Ges. 1926, 59, 1254. Zemple´n, G.;
Kiss, D. Ber. Dtsch. Chem. Ges. 1927, 60, 165.

5794 J. Am. Chem. Soc., Vol. 119, No. 25, 1997
McCarter et al.
Scheme 2
In the high stereoselectivity approach (Scheme 2) the bro-
mosugar 1 was treated with sodium ethanethiolate to give the
ethyl â-thioglycoside 15 exclusively. Replacement of the
acetate protecting groups with benzyl ethers was achieved by
treating 15 with benzyl bromide in THF in the presence of
sodium hydroxide. The newly protected thioglycoside 16 was
converted to the bromide 17 by reaction with bromine. Addition
of the partially protected anhydro sugar 3 to 17 in dichlo-
romethane in the presence of 1.2 equiv of silver triflate at -78
°C afforded a mixture of the fluorodisaccharides 18 and 19.
Chromatographic separation of the fluorodisaccharides from
hydrolyzed and unreacted starting materials gave a 10/90 ratio
of the R- and â-1,4-linked fluorodisaccharide products from
which the two anomers could be isolated in >95% purity by
rechromatography. Hydrogenolysis of the benzyl ethers of 18
and 19 gave the partially protected fluorodisaccharides 20 and
21. Acetylation of 20 and 21 gave products that were identical
to fluorodisaccharides 4 and 5, respectively.
2′-Deoxy-2′-fluorolactose (14). In a procedure identical to
that for the synthesis of 2′-deoxy-2′-fluorocellobiose (8) Via the
low selectivity method, the fluorogalactosyl bromide 2 (Scheme
1) was reacted with the anhydro sugar 3 in acetonitrile in the
presence of silver salt promoters. A mixture of the previously
reported10 2′-deoxy-2′-fluorodisaccharide derivative 9 and the
novel 2′-deoxy-2′-fluorolactose derivative 10 was chromato-
graphically isolated along with the unreacted anhydro sugar 3
and the hydrolysis product 11. The fluorodisaccharides were
isolated in a ∼60/40 ratio of R-linked to â-linked anomers and
the R-linked anomer was crystallized from the mixture. How-
ever, 2′-deoxy-2′-fluorolactose derivative 10 in the mother liquor
could not be induced to crystallize despite its predominance
(90%), nor could this enriched mixture be further purified by
chromatography. In order to obtain a pure sample of 10 the
mixture of anomers was treated with acetic anhydride and
sulfuric acid (acetolysis of the 1,6-anhydro ring) to give an
anomeric mixture of the fully acetylated R- and â-1,4-linked
fluorodisaccharides. The R- and â-1,4-linked fluorodisaccha-
rides could be partially separated chromatographically to give
an anomeric mixture of the 1-O-acetates of compound 12 in
high purity. This purified anomeric mixture of acetates could
not be crystallized, but transformation of the mixture to the
corresponding protected 2′-deoxy-2′-fluorolactosyl bromide 13
and treatment with mercuric acetate give exclusively the
â-acetate 12 that crystallized from ether. Deprotection of 12
with methoxide gave the syrupy 2′-deoxy-2′-fluorolactose (14).
Figure 1. Inactivation of Agrobacterium â-glucosidase with 8. Enzyme
was incubated with the following concentrations of 8: 0.57 mM (4),
1.30 mM (9), 2.60 mM (0), 4.90 mM (b), 9.80 mM ([), 20.0 mM
(2). (a) Residual activity versus time. (b) Replot of rate constants from
(a). (c) Protection against inactivation given by isopropyl â-D-thio-
glucoside: (O) 3.27 mM 8 alone, (b) 3.27 mM 8 plus isopropyl â-D-
thioglucoside (8 mM).
Specific Inactivation of a Glycosidase by a 2′-Deoxy-2′-
fluorodisaccharide. 2′-Deoxy-2′-fluorocellobiose (8) inacti-
vated Agrobacterium â-glucosidase in a time-dependent fashion
(ki ) 0.018 ( 0.003 min-1, Ki ) 20 ( 6 mM), presumably Via
the accumulation of a 2-deoxy-2-fluoroglucosyl-enzyme inter-
mediate (Figure 1).12 Evidence for this mechanism was
accumulated in a number of ways. Firstly, inactivation was
shown not to be due to a contaminating impurity, such as an
activated fluorosugar not observed by NMR, by incubating a
high concentration of â-glucosidase (0.55 mM) with 8 (1.04
mM) and monitoring loss of activity. Full inactivation was
observed on a time scale comparable to that when much lower
enzyme concentrations were employed, indicating that the
inactivating species must be 8 since were it to be a contaminant
present at low levels then there would not have been a sufficient
amount to inactivate this concentration of enzyme. Secondly,
8 was shown to be active site-directed since protection against
inactivation was afforded by the competitive inhibitor isopropyl
â-D-thioglucoside (8 mM, Ki ) 4 mM),4 decreasing kobs at 3.27
mM inactivator from 0.0024 to 0.00022 min-1. Thirdly, the
formation of a 2-deoxy-2-fluoro-R-glucosyl-enzyme intermedi-
ate was indicated by its reaction behavior which was identical
to that of the previously characterized species. Thus, freed of
excess 8 and incubated in the presence of â-D-glucosylbenzene
(100 mM) which acts as a transglucosylation acceptor, reactiva-
tion of the isolated 2-deoxy-2-fluoroglycosyl-enzyme was
observed, with a kreac of (2.9 ( 0.5) × 10-3 min-1, comparable
to that seen previously (3.3 × 10-3 min-1) with the 2-deoxy-
2-fluoroglucosyl-enzyme13 (produced by reaction with 2-deoxy-
2-fluoro-â-D-glucosyl fluoride) at the same concentration of â-D-
glucosylbenzene. Further, electrospray mass spectrometry
(12) Kinetic parameters for inactivation were determined essentially as
reported previously4,13 using inactivator concentrations of 0.57, 1.30, 2.60,
4.90, 9.80, and 20.00 mM.
(13) Street, I. P.; Kempton, J. B.; Withers, S. G. Biochemistry 1992, 31,
9970.

Design and Synthesis of 2′-Deoxy-2′-fluorodisaccharides
J. Am. Chem. Soc., Vol. 119, No. 25, 1997 5795
indicated that the mass of the inactivated enzyme was increased
from 51 192 ( 6 to 51 354 ( 5, consistent, within error, with
the added mass of the expected 2-deoxy-2-fluoroglucosyl moiety
(165). Finally, peptic digestion of the inactivated enzyme and
anlysis of the fragment peptides by HPLC-tandem mass
spectrometry14 showed that the inhibitor was covalently bound
to two peptides with retention times and masses identical to
those of the YITENGAC (m/z ) 1035) and ITENGAC (m/z )
871) peptides previously labeled by the activated inhibitor 2′,4′-
dinitrophenyl-2-deoxy-2-fluoro-â-D-glucoside, indicating that the
same nucleophilic residue, previously identified as Glu-358,15,21
was labeled by both inhibitors
the rate of the glycosylation step for the intermediate to
accumulate upon reaction of 2′-deoxy-2′-fluorocellobiose with
this enzyme. By contrast, reaction of this compound with
Agrobacterium â-glucosidase, which specifically binds glucose
in its aglycon subsite, results in a tremendous acceleration of
the glycosylation step, and inactivation of the enzyme. Such
specificity may be useful in the design of stable, highly specific
inhibitors of glycosidases (such as those involved in glycoprotein
processing) whose differences in specificity are based principally
upon differences in substrate aglycon structure.
Experimental Section
In keeping with expectations, 2′-deoxy-2′-fluorolactose (14)
was not an inactivator of the â-galactosidase from Escherichia
coli. The kcat of lactose, the natural substrate of this enzyme,
is 2.5-fold lower than that of p-nitrophenyl â-D-galactoside, a
substrate for which galactosylation is rate-limiting.16 Therefore,
accumulation of a 2-deoxy-2-fluorogalactosyl-enzyme derived
from 2′-deoxy-2′-fluorolactose was not expected since this
compound would similarly be expected to exhibit rate-limiting
galactosylation. Indeed, incubation of â-galactosidase with 14
for 1 h at concentrations up to 8.9 mM did not result in
inactivation. Instead, 14 was an exceedingly slow substrate of
this enzyme, functioning effectively as a competitive inhibitor
(Ki ) 0.9 mM). This result contrasts sharply with the rapid
inactivation (t1/2 ≈ 3 s at saturation) of E. coli â-galactosidase
General Methods. Anhydrous solvents were prepared according
to literature22 procedures except for diethyl ether which was used
directly as anhydrous reagent grade. Silica gel used for column
chromatography was Merck silica gel 60, and elution was performed
at ∼8 psi above atmospheric pressure. The acetylated glucal and
galactal were obtained from Terochem Laboratories, and the trifluo-
romethyl hypofluorite was from PCR. 2,3-Di-O-acetyl-1,6-anhydro-
â-D-glucose was prepared according to literature23 procedures. Melting
points were measured on a Leitz 350 microscope heating stage and are
uncorrected. Optical rotations were measured in a 1.0 dm tube with a
Perkin-Elmer 141 polarimeter. 1H-NMR spectra were recorded with a
Bruker WH-400 spectrometer, using CDCl3 as solvent, unless stated
otherwise. Mass spectra were determined on a Varian/MAT CH4B or
a Kratos MS50 spectrometer. Elemental analyses were performed at
1
UBC by Mr. P. Borda. Ratios of isomers reported are based on H-
NMR integration.
by 2-deoxy-2-fluoro-â-D-galactosyl fluoride (ki ) 13.2 min-1
,
Ki ) 1.3 mM),5 wherein the fluoride leaving group serves to
accelerate the galactosylation step sufficiently such that dega-
lactosylation becomes rate-limiting. However, lactose-hydro-
lyzing enzymes for which degalactosylation is rate-limiting with
the natural substrate would be predicted to be inactivated by
14.
3,4,6-Tri-O-acetyl-2-deoxy-2-fluoro-r-D-gluco(and galacto)pyra-
nosyl bromide10 (1 and 2). The bromides 1 and 2 were prepared as
described in the literature10 from 3,4,6-tri-O-acetyl-D-glucal (or galactal)
and trifluoromethyl hypofluorite in Freon-11. The chromatographically
isolated 2-deoxy-2-fluoro-R-D-gluco(or galacto)pyranosyl fluoride was
treated with HBr/AcOH, and evaporation of the volatiles followed by
coevaporation with toluene (3×) gave crude 1 or 2 which was used in
the glycosylation reactions without further purification.
The aglycon specificity of this class of inactivators was
demonstrated by comparing two â-glucosidases with different
aglycon specificities. Glucocerebrosidase (GCase) cleaves the
â-glucosidic bond in glucosyl ceramide to â-D-glucose and
ceramide (N-acylsphingosine) and is an essential catabolic
enzyme in the lysosomal compartment of mammalian cells. Like
Agrobacterium â-glucosidase, human GCase is effectively
inactivated by 2-deoxy-2-fluoro-â-D-glucosyl fluoride (ki/Ki )
0.023 min-1 mM-1), an inactivator in which the fluoride leaving
group serves as an activated, but nonspecific, aglycon with both
enzymes.17 Significantly, no detectable inactivation of GCase
by 2′-deoxy-2′-fluorocellobiose (8) after 24 h of incubation at
concentrations up to 2.2 mM was observed. Evidently, and quite
reasonably, binding interactions in the aglycon subsite of GCase
which have presumably evolved to specifically bind the N-
acylsphingosyl moiety, do not result in a sufficient increase in
2′-Deoxy-2′-fluorocellobiose (8). Glycosylation in Acetonitrile.
2,3-Di-O-acetyl-1,6-anhydro-4-O-(3,4,6-tri-O-acetyl-2-deoxy-2-fluoro-
â-D-glucopyranosyl)-r-(and â-)D-glucopyranose (4 and 5). A solu-
tion of bromide 1 (0.42 g, 1.35 mmol) in absolute ether (0.8 mL) was
added slowly to a stirred solution of the anhydro sugar 3 (3.5 g, 14
mmol) in anhydrous acetonitrile (10 mL) containing silver trifluo-
romethanesulfonate (CF3SO3Ag) (0.307 g, 1.2 mmol), silver carbonate
(Ag2CO3) (0.817 g, 3 mmol), and powdered Drierite (0.70 g). The
resulting mixture was stirred overnight at room temperature and filtered
through Celite. The residue was washed with CH2Cl2. The combined
filtrates were evaporated and the products separated on a silica gel
column (1:2 (v/v) hexanes-ethyl acetate) to give the hydrolysis product
6, a 70:30 mixture of 4 and 5 (300 mg; 41%), and excess anhydro
sugar 3. The fluoromaltose derivative 4 was crystallized in ether from
the above mixture of 4 and 5, and the supernatant was chromatographed
on a silica gel column using ether as an eluent. The fractions containing
>50% of the more mobile 5 were pooled, and 5 was crystallized from
ether and recrystallized from ethanol to a constant melting point. Mp
(14) The labeled peptides were identified using tandem mass spectrometry
in neutral loss mode, in which the ions are subjected to limited fragmentation
by an inert gas in a collision cell.19 The ester bond between inhibitor and
peptide is one of the more labile bonds present, undergoing homolytic
cleavage with loss of a neutral sugar. The two quadrupoles are scanned in
a linked mode so that only those ions differing by the mass of the label
could be detected. For a singly-charged peptide, the m/z difference is the
mass of the label (165); for a doubly-charged peptide, the m/z difference is
half of the mass of the label, and so on. See refs 19 and 20.
(15) Withers, S. G.; Warren, R. A. J.; Street, I. P.; Rupitz, K.; Kempton,
J. B.; Aebersold, R. J. Am. Chem. Soc. 1990, 112, 5887.
(16) Rosenberg, S.; Kirsch, J. F. Biochemistry 1981, 20, 3189.
(17) Miao, S.; McCarter, J. D.; Grace, M.; Grabowski, G.; Aebersold,
R.; Withers, S. G. J. Biol. Chem. 1994, 269, 10975.
(18) Sinnott, M. L. Chem. ReV. 1990, 90, 1171.
(19) Busch, K. L.; Cooks, R. G. Tandem Mass Spectrometry; John Wiley
& Sons: New York, 1983.
(20) Miao, S.; Ziser, L.; Aebersold, R.; Withers, S. G. Biochemistry 1994,
33, 7027.
(21) Tull, D.; Miao, S.; Withers, S. G.; Aebersold, R. Anal. Biochem.
1995, 224, 509.
104-107 °C. [R]22 -28.2° (c 0.86, CHCl3). 1H-NMR: δ 2.04 (s,
D
3H × 2, OAc), 2.10, 2.12, 2.14 (each s, 3H × 3, OAc), 3.51 (s, H-4),
3.79 (dd, Jgem ) 7.5 Hz, J5,6a ) 5.5 Hz, H-6a), 3.87 (ddd, J4′,5′ ) 9.5
Hz, J5′,6′a ) 2.0 Hz, J5′,6′b ) 5.5 Hz, H-5′), 3.90 (d, H-6b), 4.10 (dd,
Jgem ) 12 Hz, H-6′a), 4.24 (dd, H-6′b), 4.41 (ddd, J1′,2′ ) 8.0 Hz, J2′,F
) 51 Hz, J2′,3′ ) 9.0 Hz, H-2′), 4.51 (s, H-2), 4.66 (d, H-5), 4.94 (dd,
J1′,F ) 3.0 Hz, H-1′), 4.99 (dd, J3′,4′ ) 9.0 Hz, H-4′), 5.23 (s, H-3),
5.35 (ddd, JF,3′ ) 14.5 Hz, H-3′), 5.42 (s, H-1). MS (EI): m/z 537 (M
+ H), 477 (M - CH3CO2), 463 (M - CH3CO3CH2), 291 (M -
aglycon) Anal. Calcd for C22H29FO14: C, 49.25; H, 5.45. Found: C,
49.30; H, 5.45.
(22) Perrin, D. D.; Armarego, W. L. F.; Perrin, D. R. Purification of
Laboratory Chemicals; Pergamon Press: New York, 1980.
(23) Shapiro, D.; Rabinsohn, Y.; Acher, A. J.; Diver-Haber, A. J. Org.
Chem. 1970, 35, 1464. Ward, R. B. Methods in Carbohydrate Chemistry;
Academic Press: New York, 1962; p 394.

5796 J. Am. Chem. Soc., Vol. 119, No. 25, 1997
McCarter et al.
1,2,3,6-Tetra-O-acetyl-4-O-(3,4,6-tri-O-acetyl-2-deoxy-2-fluoro-
â-D-glucopyranosyl)-r-(and â-)D-glucopyranose (7). Acetolysis of
the fluorocellobiose derivative 5 (626 mg, 1.2 mmol) gave an oil which
could not be induced to crystallize. Purification of the oil by column
chromatography with 19:1 benzene-ethanol gave 609 mg (0.954 mmol,
82%) of an oil consisting of a mixture of the anomeric acetates of 7
(2:1 R/â) which again could not be induced to crystallize. Careful
rechromatography using the above conditions and isolation of the early
and later fractions gave samples of ∼90% anomeric purity from which
1H-NMR spectra and optical rotations were obtained. ∼90% R-acetate.
[R]22D +86.3° (c 1.2, CHCl3). 1H-NMR: δ 2.00, 2.01, 2.03, 2.05, 2.07,
was filtered through Celite and the residue washed thoroughly with
dichloromethane. The filtrate was evaporated in Vacuo and the residue
dried at 250 °C under high vacuum. The syrup was purified on a
column of silica gel (300 g, hexanes-ether, 4:1) to give syrupy 16
(3.6 g, 86%). 1H-NMR: δ 1.31 (t, J ) 8 Hz, CH2CH3), 2.74 (m, CH2-
CH3), 3.47 (ddd, J4,5 ) 10 Hz, J5,6a ) 4.5 Hz, J5,6b ) 2 Hz, H-5), 3.62
(dd, J3,4 ) 9 Hz, H-4), 3.66 (dd, Jgem ) 11.5 Hz, H-6a), 3.73 (dd,
H-6b), 3.76 (ddd, J2,3 ) 8.5 Hz, JF,3 ) 15 Hz, H-3), 4.35 (ddd, J1,2
)
10 Hz, J2,F ) 49.5 Hz, H-2), 4.48 (dd, J1,F ) 2 Hz, H-1), 4.54 (×2),
4.575, 4.74, 4.83, 4.91 (each d, 1H × 6, Jgem - 11 Hz, CH2Ph), 7.15-
7.37 (m, 15 H, ArH). MS (EI): m/z 476 (M - HF), 405 (M - PhCH2).
2.10, 2.15 (each s, 3H × 7, OAc), 3.69 (ddd, J4′,5′ ) 9.9 Hz, J5′,6′a
)
3,4,6-Tri-O-benzyl-2-deoxy-2-fluoro-r-D-glucopyranosyl Bromide
(17). To a stirred solution of the benzylated ethyl thioglycoside 16
(940 mg) in anhydrous ethyl ether (15 mL) was added, dropwise, a
solution of bromine (0.15 mL in 3 mL of ether). After 1 h the solution
was coevaporated with toluene (2 × 4 mL) and the residue dried in
Vacuo to give the benzylated fluoroglucosyl bromide 17 as a moderately
viscous amber syrup which was used in the glycosylation reaction
without further purification.
2.2 Hz, J5′,6′b ) 4.3 Hz, H-5′), 3.86 (dd, J3,4 ) 9.2 Hz, J4,5 ) 10.1 Hz,
H-4), 4.04 (dd, Jgem ) 12.5 Hz, H-6′a), 4.07 (dd, J5,6a ) 2.5, H-5),
4.19 (dd, Jgem ) 12.4 Hz, H-6a), 4.25 (ddd, J1′,2′ ) 7.7 Hz, J2′,F ) 48.5
Hz, J2′,3′ ) 8.7 Hz, H-2′), 4.34 (dd, H-6′b), 4.48 (d, H-6b), 4.59 (dd,
J1′,F ) 2.7, H-2′), 4.98 (dd, J3′,4′ ) 9.7 Hz, H-4′), 5.01 (dd, J1,2 ) 3.7
Hz, J2,3 ) 10.3 Hz, H-2), 5.27 (ddd, JF,3′ ) 14.7 Hz, H-3′), 5.49 (dd,
H-3), 6.26 (d, H-1). Anal. Calcd for C26H35FO7: C, 48.90; H, 5.52.
Found: C, 48.63; H, 5.52. ∼90 â-acetate. [R]22 +35.6° (c 0.80,
D
2,3-Di-O-acetyl-1,6-anhydro-4-O-(3,4,6-tri-O-benzyl-2-deoxy-2-
fluoro-r-(and â-D-glucopyranosyl)-â-D-glucopyranose (18 and 19).
The procedure of Kronzer and Schuerch24 was modified for the
following glycosylation. Silver trifluoromethanesulfonate (233 mg, 0.91
mmol) was dried in the reaction flask for several hours in Vacuo, the
flask was cooled to -78 °C, and a solution of bromide 17 (300 mg in
2 mL of CH2Cl2) was slowly added to the stirred mixture. After 10
min a solution of the anhydro sugar 3 (200 mg in 2 mL of CH2Cl2)
was slowly added to the above mixture. After 30 min at -78 °C the
mixture was allowed to warm under ambient conditions for 10 min.
The reaction was terminated by shaking with excess saturated sodium
hydrogen carbonate. The mixture was filtered through Celite and the
organic layer washed with water, dried over sodium sulfate, filtered,
and evaporated to a syrup. The syrup was purified by column
chromatography (hexanes-ethyl acetate, 3:2) to give 248 mg (62.5%)
of a 9:1 mixture of the â-linked to R-linked disaccharides 19 and 18.
Careful rechromatography (benzene-ethyl acetate, 4:1) gave 18 and
19 each in >95% anomeric purity, with the â-linked 19 as the more
mobile component. 1H-NMR (18): δ 2.05, 2.09 (each s, 3H × 2, OAc),
3.57 (ddd, J ) 2, 3.5, 9.5 Hz, H-5′), 3.60 (br s, H-4), 3.70-3.81 (m,
4 H, H-3′, H-4′, H-6′a, H-6′b), 3.82 (dd, Jgem ) 7.7 Hz, J5,6 ) 5.5 Hz,
H-6a), 3.98 (d, H-6b), 4.455 (ddd, J1′,2′ ) 8 Hz, J2′,3′ ) 8 Hz, J2′,F ) 51
Hz, H-2′), 4.59 (br s, H-2), 4.73 (d, H-5), 4.77 (dd, J1′,F ) 3.3 Hz,
H-1′), 4.465, 4.57, 4.61, 4.76, 4.85, 4.91 (each d, 1H, × 6, Jgem - 11
Hz, CH2Ph), 5.35 (dd, J2,3 ) 1.5 Hz, J3,4 ) 1.5 Hz, H-3), 5.49 (s, H-1),
7.13-7.35 (m, 15 H, ArH). 1H-NMR (19): δ 2.08, 2.09 (each s, 3H
× 2, OAc), 3.55 (br s, H-4), 3.61 (dd, J3′,4′ ) 9.0 Hz, J4′,5′ ) 10.5 Hz,
H-4′), 3.67 (ddd, J5′,6′a ) 2.5 Hz, Jgem ) 10.5 Hz, H-6′a), 3.68 (dd,
J5′,6′b ) 4.0 Hz, H-6′b), 3.71 (dd, J5,6a ) 5.5 Hz, Jgem ) 8.0 Hz, H-6a),
3.96 (d, J5,6b ) 0.5 Hz, H-6b), 4.04 (ddd, H-5′), 4.15 (ddd, J2′,3′ ) 9.0
Hz, JF,3′ ) 12.2 Hz, H-3′), 4.51 (ddd, J1′,2′ ) 4.0 Hz, J2′,F ) 50 Hz,
H-2′), 4.58 (br s, H-2), 4.66 (br d, H-5), 4.835 (dd, J2,3 ) 1.5 Hz, J3,4
) 1.5 Hz, H-3), 4.455, 4.475, 4.555, 4.735, 4.83, 4.87 (each d, 1H, ×
6, Jgem - 11 Hz, CH2Ph), 5.30 (d, H-1′), 5.44 (s, H-1), 7.145-7.41
(m, 15 H, ArH).
CHCl3). 1H-NMR: (the resonances and coupling constants of the
fluorinated nonreducing end of the â-acetate of 7 were effectively
unchanged compared to those of the R-acetate above except for J2′,F
which increased to 49.9 Hz) δ 3.79 (ddd, J4,5 ) 9.5 Hz, J5,6a ) 4.2 Hz,
J5,6b ) 1.85 Hz, H-5), 3.89 (dd, J3,4 ) 9.5 Hz, H-4), 4.20 (dd, Jgem
)
12.25 Hz, H-6a), 4.47 (dd, H-6b), 5.05 (dd, J1,2 ) 8.3 Hz, J2,3 ) 9.5
Hz, H-2), 5.27 (dd, H-3), 5.69 (d, H-1).
4-O-(2-Deoxy-2-fluoro-â-D-glucopyranosyl)-D-glucopyranose (8,
2′-Deoxy-2′-fluorocellobiose). To a syrupy mixture of the R- and
â-anomeric acetates of 7 (271 mg, 0.425 mmol) under an atmosphere
of nitrogen was added 100% methanol (10 mL). The mixture was
sonicated until all the syrup had dissolved, after which methanolic
NaOCH3 (1.7 M, 0.15 mL) was added and the mixture stirred for 1 h.
The resulting yellow solution was neutralized (as indicated on pH paper)
with Amberlite IR-120 (H+) resin and filtered and the residue washed
with methanol. The combined filtrates were concentrated in Vacuo,
treated with charcoal, and filtered through Celite to give 8 (138 mg,
94.4%) as a hygroscopic glass after evaporation of methanol and drying
under high vacuum. [R]22 +35.9° (c 0.90, H2O). 1H-NMR (D2O):
D
(splitting in H-1′ and H-2′ due to anomeric mixture on reducing glucose
residue) δ 4.17/4.175 (ddd × 2, J1′,2′ ) 7.7 Hz, J2′,F ) 51.0 Hz, J2′,3′
)
8.8 Hz, H-2′), 4.65 (d, J1,2 ) 8.0 Hz, 0.6 H, H-1 â-anomer), 4.79/
4.795 (d × 2, H-1′), 5.21 (d, J1,2 ) 3.5 Hz, 0.4 H, H-1 R-anomer).
19F-NMR (D2O, TFA ref): δ -122.45 (dd, JH2′,F ) 50.6 Hz, JH3′,F
15.3 Hz). Anal. Calcd for C12H21FO10‚(1/2)H2O: C, 40.79; H, 6.28.
Found: C, 40.66; H, 6.46.
)
Glycosylation with Benzyl-Protected Glucosyl Bromide at Low
Temperature. Ethyl 3,4,6-Tri-O-acetyl-2-deoxy-2-fluoro-â-D-thio-
glucopyranoside (15). A solution of the fluoroglucosyl bromide 1
(5.6 g, 15.1 mmol) in anhydrous methanol (10 mL) was treated with a
2.0 mL portion of an ethanethiolate solution (10 mL) (prepared by
dissolving potassium (700 mg, 17.9 mmol) portionwise in anhydrous
methanol (10 mL) followed by ethanethiol (2.0 mL) over 15 min. After
stirring for 0.5 h, the violet-brown solution was filtered and concentrated
in Vacuo to a syrup. The syrup was acetylated by refluxing in acetic
anhydride (35 mL) and anhydrous sodium acetate (3.5 g) for 10 min.
The cooled mixture was poured onto 150 g of crushed ice and
vigorously stirred. The precipitate was filtered, washed thoroughly with
water, and crystallized (dichloromethane-hexanes) to give 15 (4.15
g; 65%). Mp 124-126.5 °C. 1H-NMR: δ 1.34 (t, CH2CH3), 2.04,
Disaccharide 4 from Disaccharide 18 Wia Intermediate 2,3-Di-
O-acetyl-1,6-anhydro-4-O-(2-deoxy-2-fluoro-r-D-glucopyranosyl)-
â-D-glucopyranose (20). To a prehydrogenated mixture of 10%
palladium on charcoal (72 mg) and ethanol (5 mL) was added a solution
of the benzylated disaccharide 18 (60 mg, 0.106 mmol) in ethanol (5
mL). The mixture was hydrogenated at atmospheric pressure for 2.5
h, filtered, and concentrated to yield the debenzylated product 20 (34
mg, 0.083 mmol, 78%) as a colorless syrup. The crude syrup was
dissolved in acetic anhydride-pyridine (3 mL, 1:1) at 0 °C and stirred
at room temperature for 12 h, concentrated, coevaporated with toluene,
and purified by column chromatography (benzene-ethyl acetate, 1:1)
2.08, 2.09 (each s, 3H × 3, OAc), 2.76 (m, CH2CH3), 3.73 (m, J4,5
10 Hz, J5,6a ) 2.5 Hz, J5,6b ) 5.0 Hz, H-5), 4.12 (dd, Jgem ) 12.4 Hz,
H-6a), 4.24 (dd, H-6b), 4.32 (ddd, J1,2 ) 10 Hz, J2,F ) 49.5 Hz, J2,3
)
)
9 Hz, H-2), 4.60 (dd, J1,F ) 1.5, H-1), 5.03 (dd, J3,4 ) 9.5 Hz, H-4),
5.32 (m, J3,F ) 14.5 Hz, H-3). MS (EI): m/z 352 (M), 291 (M -
SEt), 272 (291 - F). Anal. Calcd for C14H21FO7S: C, 47.72; H, 6.01.
Found: C, 47.72; H, 6.03.
1
to give 4 (34 mg; 60% from 18). This compound produced a H-
NMR spectrum identical to that of compound 4 synthesized above.
Ethyl 3,4,6-Tri-O-benzyl-2-deoxy-2-fluoro-â-D-thioglucopyrano-
side (16). To a stirred solution of the acetylated ethyl thioglycoside
15 (3.0 g) in redistilled tetrahydrofuran (20 mL) was added powdered
potassium hydroxide (8.0 g) followed by dropwise addition of benzyl
bromide (10 mL). After refluxing for 15 h and cooling, the mixture
Disaccharide 5 from Disaccharide 19 Wia Intermediate 2,3-Di-
O-acetyl-1,6-anhydro-4-O-(2-deoxy-2-fluoro-r-D-glucopyranosyl)-
â-D-glucopyranose (21). Treatment of the benzylated cellobiose
(24) Kronzer, F. J.; Schuerch, C. Carbohydr. Res. 1973, 27, 379.

Design and Synthesis of 2′-Deoxy-2′-fluorodisaccharides
J. Am. Chem. Soc., Vol. 119, No. 25, 1997 5797
19F-NMR (D2O, TFA ref): δ -130.73 (ddd, JH2′,F ) 51.4 Hz, JH3′,F
derivative 19 (30 mg, 0.044 mmol) under conditions analogous to those
used for the benzylated maltose derivative 18 above gave 5 (16 mg,
66%) after chromatography. This compound produced a 1H-NMR
spectrum identical to that of compound 5 synthesized above.
)
14.1 Hz, JH1′,F ≈ 3 Hz). Anal. Calcd for C12H21FO10‚(3/4)H2O: C,
40.28; H, 6.34. Found: C, 40.21; H, 6.58.
Time-Dependent Inactivation. The kinetic parameters for the
inactivation of Agrobacterium â-glucosidase by 2′-deoxy-2′-fluorocel-
lobiose were determined as follows. The enzyme was incubated at 37
°C in 50 mM sodium phosphate buffer, 0.1% BSA, pH 6.8, with the
following inhibitor concentrations: 0.57, 1.30, 2.60, 4.90, 9.80, and
20.0 mM. Aliquots (10 µL) of these inactivation mixtures were
removed at various time intervals and diluted into assay cells (1 mL)
containing p-nitrophenyl â-D-glucoside at saturating concentration (1
mM). Inactivation was monitored until 80-90% of the enzyme was
inactivated. Pseudo-first-order rate constants (kobs) for each inactivator
concentration were calculated by fitting the residual activity Versus time
data to a single exponential equation using GraFit.25 Values of ki and
Ki were determined from these kobs values by fitting to the equation
2′-Deoxy-2′-fluorolactose (14). 2,3-Di-O-acetyl-1,6-anhydro-4-O-
(3,4,6-tri-O-acetyl-2-deoxy-2-fluoro-r-(and â-)D-galactopyranosyl)-
â-D-glucopyranose (9 and 10). To a solution of the anhydro sugar 3
(3.0 g, 12 mmol) in anhydrous acetonitrile (18 mL) was added silver
trifluoromethanesulfonate (900 mg) and silver carbonate (2.76 g). A
solution of the bromide 2 (3.0 g, 8.08 mmol) in acetonitrile (5 mL)
was slowly added to the vigorously stirred mixture. The mixture was
stirred overnight in the absence of light, diluted with dichloromethane
(25 mL), and filtered through Celite and the residue washed thoroughly
with dichloromethane. The filtrate was concentrated and the residue
separated by column chromatography (hexanes-ethyl acetate, 1:1). The
disaccharides 9 and 10 were isolated as a mixture (1.7 g, 39%, ∼55:
45 R/â) along with the hydrolysis product 11 and excess 3. The
previously synthesized R-fluorodisaccharide 9 was crystallized from
ether-hexanes. The desired fluorolactose derivative 10 remained as
an uncrystallizable syrupy mixture (∼85:15 â/R-linked disaccharides).
1H-NMR (10): δ 2.03, 2.06, 2.10 (s, 3H × 3, OAc), 2.135 (each s,
6H, OAc × 2), 3.60 (br s, H-4), 3.83 (dd, J5,6a ) 6.0 Hz, Jgem ) 8.0
Hz, H-6a), 3.98 (d, H-6b), 3.99 (m, H-6′a), 4.04 (m, H-5′), 4.145 (dd,
Jgem ) 10 Hz, J5′,6′b ) 5 Hz, H-6′b), 4.595 (d, J2,3 ) 1.5 Hz, H-2),
4.615 (ddd, J1′,2′ ) 51.5 Hz, J2′,3′ ) 10.0 Hz, H-2′), 4.74 (d, H-5), 4.895
(dd, J1′,F ) 4.0 Hz, H-1′), 5.125 (ddd, JF,3′ ) 12.5 Hz, J3′,4′ ) 4.0 Hz,
H-3′), 5.275 (dd, J3,4 ) 1.5 Hz, H-3), 5.40 (dd, JF,4′ ) 1.5 Hz, H-4′),
5.49 (s, H-1).
kobs ) ki[I]/(Ki + [I])
Protection against Inactivation. Protection against inactivation was
investigated as follows. Samples of Agrobacterium â-glucosidase were
incubated at 37 °C in 50 mM sodium phosphate buffer, 0.1% BSA,
pH 6.8, containing 2′-deoxy-2′-fluorocellobiose (3.27 mM), in the
presence or absence of a competitive inhibitor, isopropyl â-D-thioglu-
coside (8 mM). Aliquots were removed at various time intervals and
diluted into assay cells containing saturating concentrations of substrate,
and residual activity was monitored by following the release of
p-nitrophenolate at 400 nm as described above. Pseudo-first-order rate
constants for inactivation at the same inactivator concentration, but in
the absence or presence of the competitive inhibitor, were determined.
Reactivation of Inactivated Enzyme. An appropriate dilution of
the inactivated enzyme (200 µL) was concentrated at 4 °C using 10
kDa nominal cutoff centrifugal concentrators (Amicon Corp., Danvers,
MD) to a volume of approximately 15 µL and then diluted with 180
µL of buffer. This was repeated twice, and the retentate was diluted
to a final volume of buffer (165 µL) containing 1 mg/mL BSA and the
transglycosylation ligand â-D-glucosylbenzene (100 mM). The inac-
tivated enzyme was then incubated at 37 °C, and reactivation was
monitored by removal of aliquots (10 µL) at appropriate time intervals
and assaying as described above. Measured activities were corrected
for decreases in activity due to denaturation over this time course using
a noninhibited control sample. The spontaneous reactivation rate
constant, k3, was determined by fitting the data so obtained to a first-
order rate equation, as described above.
1,2,3,6-Tetra-O-acetyl-4-O-(3,4,6-tri-O-acetyl-2-deoxy-2-fluoro-
â-D-galactopyranosyl)-â-D-glucopyranose (12). The 1,6-anhydro ring
of 10 (∼1.6 g, containing ∼0.27 g of 9 as an inseparable mixture) was
acetolyzed as previously described. The resulting syrup was purified
by column chromatography (neat diethyl ether) to give an inseparable
mixture of the R- and â-anomeric acetates of 12 (859 mg, 45%) in a
1.5:1 ratio. A sample of this mixture (750 mg) was dissolved in 45%
HBr in AcOH (4.5 mL) and stirred for 0.5 h. The solution was diluted
with chloroform (15 mL), washed with water (2 × 15 mL), dried over
anhydrous sodium sulfate, filtered, and concentrated. The resulting
syrup of crude 13 was dissolved in acetic anhydride (10 mL) containing
mercuric acetate (0.75 g) and stirred for 2 h. The solution was diluted
with chloroform (30 mL), washed with water (2 × 20 mL), dried over
anhydrous sodium sulfate, concentrated, crystallized from ether, and
recrystallized from ethanol to give the pure â-1,4-linked â-O-acetate
12 (650 mg, 87%). Mp 96-103 °C (partial melt, resolidifies), 161.5-
163 °C. [R]22 +23.1 (c 1.5, CHCl3), [R]22 -4.4 in CHCl3 for the
Test for Inhibitory Contaminant in 2′-Deoxy-2′-fluorocellobiose.
Agrobacterium â-glucosidase (37 mg/mL, 29 µL, 2.2 × 10-8 mol) was
incubated with 2′-deoxy-2′-fluorocellobiose (3.78 mM, 11 µL, 4.2 ×
10-8 mol) in a total volume of 40 µL for 44 h at 37 °C. The molar
ratio of inhibitor/enzyme is thus ∼1.9:1. Aliquots were removed at
time intervals, diluted appropriately, and assayed with p-nitrophenyl
â-D-glucoside as described previously. The amount of inactivation after
∼25-45 h of incubation was compared to that previously determined
after similar incubation times, but with a much greater inhibitor excess.
D
D
2′-OAc analogue. MS (EI): m/z 618 (M - HF), 595 (M - CH3CO),
579 (M - CH3CO2), 291 (M - aglycon). 1H-NMR: δ 2.06 (s, 6 H,
OAc × 2), 2.060, 2.062, 2.098, 2.100, 2.132 (each s, 3H × 5, OAc),
3.81 (ddd, J4,5 ) 10 Hz, J5,6a ) 4 Hz, J5,6b ) 2 Hz, H-5), 3.90 (dd, J4′,5′
) 0.5 Hz, J5′,6′a ) 7 Hz, J5′,6′b ) 6.5 Hz, H-5′), 3.94 (dd, J3,4 ) 9.6 Hz,
H-4), 4.06 (dd, Jgem ) 11.5 Hz, H-6′a), 4.11 (dd, H-6′b), 4.25 (dd, Jgem
) 12.5 Hz, H-6a), 4.44 (ddd, J1′,2′ ) 7.5 Hz, J2′,F ) 50.5 Hz, J2′,3′ ) 10
Hz, H-2′), 4.46 (dd, H-6b), 4.59 (dd, J1′,F ) 4.5 Hz, H-1′), 5.05 (ddd,
Acknowledgment. We thank Dr. David Burgoyne for
assistance with mass spectrometry and Dr. Marie Grace of the
Mt. Sinai Medical Center, New York, for a gift of human
glucocerebrosidase. We thank the Natural Sciences and Engi-
neering Research Council of Canada and the Protein Engineering
Network of the Centres of Excellence of Canada for financial
support.
J3′,4′ ) 3.7 Hz, JF,3′ ) 12.7 Hz, H-3′), 5.055 (dd, J1,2 ) 8.4 Hz, J2,3
)
9.6 Hz, H-2), 5.295 (dd, H-3), 5.38 (br dd, JF,4′ ) 3.7 Hz, H-4′), 5.705
(d, H-1). Anal. Calcd for C26H35FO17‚(1/2)H2O: C, 48.22; H, 5.60.
Found: C, 48.54; H, 5.73.
4-O-(2-Deoxy-2-fluoro-â-D-galactopyranosyl)-D-glucose (2′-Deoxy-
2′-fluorolactose, 14). The hepta-O-acetate 12 (54 mg, 0.085 mmol)
was treated in a manner identical to that for the production of 8 from
7 to give 14 (27 mg, 88%) as a hygroscopic transparent glass after
pumping at high vacuum. [R]22 +53.7° (c 1.2, H2O), [R]22 +55.3°
JA9627454
D
D
(H2O) for the 2′-hydroxy analogue. 1H-NMR (D2O): δ 5.22 (d, J1,2
)
(25) Leatherbarrow, R. J. GraFit Version 3.0, Erithacus Software Ltd.,
Staines, U.K., 1990.
4.0 Hz, H-1 (R-anomer)), 4.38 (m, H-2′ (R-anomer), H-2′ (â-anomer)).