Syntheses of methyl (4,6-dideoxy-α-L-lyxo-hexopyranosyl)-(1→3)- and (4-deoxy-4-fluoro-α-L-rhamnopyranosyl)-(1→3)-2-acetamido-2-deoxy-α-D-glucopyranosides, analogs of the mycobacterial arabinogalactan linkage disaccharide

Article abstract of DOI:10.1016/S0008-6215(99)00203-7

We have made thioglycoside donors for the 4,6-dideoxy-L-lyxo-hexopyranosyl ('4-deoxy-L-rhamnosyl') and 4-deoxy-4-fluoro-L-rhamnosyl monosaccharide residues. The preparation of the deoxyfluororhamnose was not straightforward, and revealed some unexpected behavior of the diethylaminosulfur trifluoride (DAST) reagent. The new glycosyl donors were used to synthesize two analogs of the mycobacterial arabinogalactan linkage disaccharide →4)-α-L-Rha-(1→3)-α-D-GlcNAc. These analogs are prototypes for a family of potential inhibitors of the enzymes involved in the early stages of cell-wall construction in mycobacteria. Copyright (C) 1999 Elsevier Science Ltd.

Full text of DOI:10.1016/S0008-6215(99)00203-7

www.elsevier.nl/locate/carres
Carbohydrate Research 322 (1999) 14–25
Syntheses of methyl
(4,6-dideoxy-a-
L
-lyxo-hexopyranosyl)-(1“3)- and
-rhamnopyranosyl)-(1“3)-
-glucopyranosides, analogs of
(4-deoxy-4-fluoro-a-
L
2-acetamido-2-deoxy-a-
D
the mycobacterial arabinogalactan linkage disaccharide
Philip G. Hultin *, Ryan M. Buffie
Department of Chemistry, Uni6ersity of Manitoba, Winnipeg, Man., Canada R3T 2N2
Received 3 May 1999; accepted 7 July 1999
Abstract
We have made thioglycoside donors for the 4,6-dideoxy-L-lyxo-hexopyranosyl (‘4-deoxy-L-rhamnosyl’) and
4-deoxy-4-fluoro- -rhamnosyl monosaccharide residues. The preparation of the deoxyfluororhamnose was not
L
straightforward, and revealed some unexpected behavior of the diethylaminosulfur trifluoride (DAST) reagent. The
new glycosyl donors were used to synthesize two analogs of the mycobacterial arabinogalactan linkage disaccharide
“4)-a-
L
-Rha-(1“3)-a-D-GlcNAc. These analogs are prototypes for a family of potential inhibitors of the enzymes
involved in the early stages of cell-wall construction in mycobacteria. © 1999 Elsevier Science Ltd. All rights
reserved.
Keywords: Mycobacteria; Arabinogalactan; Deoxy sugars; Fluoro sugars; Diethylaminosulfur trifluoride (DAST); Thioglycosides
renewed interest in anti-mycobacterial drug
development.
1. Introduction
Mycobacteria are Gram-positive bacteria
resistant to many antibiotics, due to their
unique, carbohydrate-rich cell wall structure
[1]. A key feature of the cell-wall is an under-
lying cross-linked peptidoglycan layer sup-
porting a superstructure of arabinogalactan
(AG) [2,3]. The AG may be visualized as a
Mycobacteria cause tuberculosis and lep-
rosy, as well as a host of serious infections in
immunosuppressed people. Although many
people in the industrialized nations have come
to consider tuberculosis as a disease of the
past, it again poses a major threat. An esti-
mated 1.7 billion people carry M. tuberculosis,
leading to 8–10 million new cases of disease
and 3 million deaths annually. Antibiotic-re-
sistant and highly pathogenic strains of M.
tuberculosis have been appearing, prompting
tree whose trunk is a linear chain of
tofuranosyl residues, while its branches are
formed from -arabinofuranosyl units. Each
D
-galac-
D
branch ends in a hexa-arabinoside esterified at
its termini by long-chain mycolic acids, form-
ing a waxy outer layer.
In 1990, Brennan and co-workers revealed
that the root of the AG tree is a disaccharide
* Corresponding author. Tel.: +1-204-474-9321; fax: +1-
204-275-0905.
0008-6215/99/$ - see front matter © 1999 Elsevier Science Ltd. All rights reserved.
PII: S0008-6215(99)00203-7

P.G. Hultin, R.M. Buffie / Carbohydrate Research 322 (1999) 14–25
15
phosphate, “4)-a-
L
-Rha-(1“3)- -GlcNAc-
D
(1“P)- [4], anchoring a mycolyl-AG assem-
bly to the 6-position of each of about
10–12% of the muramic acid residues in the
peptidoglycan layer [1]. A subsequent study
[5] suggested that the biosynthesis of the my-
cobacterial cell wall begins with the synthesis
of the linkage disaccharide on a polyprenyl
pyrophosphate carrier to form 1.
Scheme 1.
was selectively acylated using the method de-
scribed by Ando et al. for the corresponding
methyl thioglycoside [12], giving the 2,3-di-O-
benzoyl derivative 3 (72%). Deoxygenation of
3 by treatment of thiocarbonylimidazolide 4
with tributyltin hydride in boiling toluene
cleanly afforded the thioglycoside donor 5 in
81% overall yield.
The galactan chain is then built up from its
4%-position. Once it is complete, the mycolyl-
AG-linker structure is finally joined to the
peptidoglycan. Although several authors have
pointed out that the linkage disaccharide is an
attractive target for rational anti-mycobacte-
rial drug design [4,6], there has been remark-
ably little work in this direction reported in
the literature to date.
The synthesis of a 4-deoxy-4-fluoro- -
rhamnosyl donor proved to be much more
challenging.
L
We are developing inhibitors for the (cur-
rently unknown) galactofuranosyl trans-
ferase(s) that build the AG onto the
Model studies were carried out on 2,3-O-iso-
propylidene (6) [13] and 2,3-di-O-benzoyl-6-
4%-position of the
L
-rhamnose in 1. Among
deoxy-a- -talopyranosides (8, obtained in low
L
the sterically conservative modifications of the
4% hydroxyl that might yield inhibitors, we
chose to begin by preparing 4%-deoxy- and
4%-deoxy-4%-fluoro- versions of the linkage dis-
accharide. This required the synthesis of suit-
able glycosyl donors for the modified
rhamnose residues, which have not been re-
ported in the literature. In this paper, we
describe the synthesis of these donors and the
preparation and characterization of two pro-
totypical modified linkage disaccharides as
their methyl glycosides.
yield by Swern oxidation/borohydride reduc-
tion of 3). NMR analysis of the products
showed that reaction of the free 4-OH groups
of 6 or 8 with diethylaminosulfur trifluoride
(DAST) in dichloromethane led to complex
mixtures. Morpholinosulfur trifluoride (mor-
pho-DAST) reacted sluggishly with 8, produc-
ing a low yield of epimeric morpholine
adducts at the 4-position. Attempts to dis-
place the 4-O-triflate from 7 [14] using KF/
18-crown-6/DMF likewise gave a mixture
whose NMR spectrum showed no evidence of
fluorine incorporation, while treatment of 9
with tris(dimethylamino)sulfur (trimethylsi-
lyl)difluoride (TASF) [15] led only to a
product whose NMR spectra suggested struc-
ture 10. These negative results are perhaps
not surprising, given the known difficulty of
nucleophilic displacements at C-4 in
monosaccharides having an axial substituent
at C-2 [16], although more potent nucle-
ophiles than F− have been successful in other
cases [14,17].
2. Results and discussion
4,6-Dideoxy
D
- and
L
-lyxo-hexopyranosyl
compounds have been prepared several times
[7–10], but no glycosyl donors for either
enantiomer of this residue have been re-
ported. We extended earlier reports to pre-
pare a thioglycoside derivative (Scheme 1).
Thus, ethyl 1-thio-a- -rhamnoside (2) [11]
L

16
P.G. Hultin, R.M. Buffie / Carbohydrate Research 322 (1999) 14–25
Scheme 2.
Since the axial C-2 substituent of the
L
-
leave the reaction overnight. This three-step,
one-pot sequence provided a 95% yield of 13
after chromatography. In contrast, when ac-
etate 19 was subjected to this sequence, we
observed that the conditions of stannylene
acetal formation promoted migration and
cleavage of the acetate group, leading to 20
(45%) and a chromatographically inseparable
1.7:1.0 mixture of 21 and 22 (11%), which
were identified by NMR spectroscopy.
talosides was evidently a problem, we turned
to a route from -fucose, whose C-2 hydroxyl
L
is equatorial. There are precedents for the
DAST fluorination of methyl 2,3-di-O-ben-
zoyl-1-thio-b-
L
-fucopyranoside [12] and 1,2,3-
-fucopyranose [18]. This
tri-O-benzoyl-a-
L
approach required inverting the C-2 stereo-
chemistry after introducing the C-4 fluorine
substituent, and to do this we needed a thiofu-
coside in which the 2- and 3-positions were
orthogonally protected. A convenient ap-
proach was suggested by the work of van
Boom and co-workers [19]. Thus, ethyl 2,3,4-
tri-O-acetyl-1-thio-b-
L
-fucopyranoside
(11)
[20] was deprotected under Zemple´n condi-
tions, and the resulting triol was immediately
converted to its 3,4-O-isopropylidene deriva-
tive. After quenching the reaction with Et3N,
the solvents were evaporated, and the crude
residue was dissolved in pyridine and treated
with benzoyl chloride. The reaction was
worked up at this stage to give 86% of crys-
talline ethyl 2-O-benzoyl-3,4-O-isopropyli-
Replacing the axial 4-OH group of 13 with
an equatorial fluorine in good yield was not
straightforward. We examined the reactions of
the 4-O-triflate 23 with TASF in dichloro-
methane or with anhydrous tetrabutylammo-
nium fluoride (TBAF) in N,N-dimethylfor-
mamide at 40 °C. In both cases a 37% yield of
the desired fluoride 14 was isolated, but the
use of TASF also led to about 10% of 24,
while TBAF formed 7% of 24 and 6% of 25
(identified by their H and C NMR spectra).
The significant proportions of elimination
products obtained were particularly trouble-
some, because in all chromatography systems
we tested 25 ran very slightly faster than 14,
while 24 ran slightly slower. As a conse-
quence, we found it impossible to completely
separate 14 from 24 by column chromatogra-
phy. The reactions of 23 with TBAF in N,N-
dimethylformamide or acetonitrile at room
temperature did not proceed to comple-
dene-1-thio-b- -fucopyranoside (12), after
L
chromatography. We also made the 2-O-
acetyl derivative 19 (71%) by a variation of
this route, using acetic anhydride in place of
benzoyl chloride at the last step (Scheme 2).
The O-isopropylidene derivative 12 was de-
protected in 80% acetic acid at 50 °C. After
evaporating the solvents in vacuo, the residual
2-O-benzoyl-3,4-diol was dried and treated
with dibutyltin oxide in boiling toluene, with
azeotropic removal of water. Evaporation of
the solvents afforded the crude 3,4-O-stan-
nylene acetal, which was then selectively ben-
zylated at O-3. The benzylation occurred quite
rapidly, but it was generally convenient to
1
13

P.G. Hultin, R.M. Buffie / Carbohydrate Research 322 (1999) 14–25
17
served were presumably this ‘activated’
intermediate. Mange and Middleton have sug-
gested [21] that displacement of the activated
intermediate by fluoride occurs concurrently
with or very soon after its formation, while
the fluoride ion and the intermediate are still
together within the solvent cage. In our case,
it would appear that because the displacement
by fluoride is sluggish, the nucleophile is able
to diffuse away. In solvents of low polarity,
extending the reaction time had little effect
because the intermediate precipitated. In more
polar solvents where the intermediate re-
mained in solution, the fluoride nucleophile
may have been more solvated, thus attenuat-
ing its nucleophilicity.
This is not the first time that an unreactive
intermediate has been observed in the DAST
reaction of a carbohydrate, but most com-
monly this situation arises when nucleophilic
attack is sterically hindered [22]. It is unlikely
that steric effects were important in our case,
as the nucleophile was approaching from the
equatorial direction. A similar situation was
reported by Yang et al. [23], who treated a
myo-inositol derivative with DAST in
dichloromethane, and recovered intact start-
ing material despite the formation of the ex-
pected sulfur ester intermediate.
tion even after 2 days, and high proportions
of 24 and 25 were also formed.
Ando et al. had obtained methyl 2,3-di-O-
benzoyl-4,6-dideoxy-4-fluoro-1-thio-b- -glu-
L
copyranoside in 45% yield from the reaction
of the corresponding fucoside with DAST in
dichloromethane [12]. When similar condi-
tions were applied to 13, we obtained 32–35%
of 14, accompanied by smaller proportions of
24 and 25 than we had observed in the reac-
tions of 23. Considerable effort was expended
to try to enhance this yield, by varying the
reagent (DAST, morpho-DAST) and the
number of equivalents used, the concentration
(0.05 M, 0.15 M, neat) and the solvent
(dichloromethane, tetrahydrofuran, toluene),
the temperature (−30 °C, 0 °C, room temper-
ature, 40 °C, 80 °C), the presence of additives
(DMAP, TASF), the type of reaction vessel
(Pyrex, Teflon-FEP), and the reaction time
(up to 5 days).
In the course of these experiments we made
several salient observations. The alcohol 13
disappeared from the TLC almost immedi-
ately after the addition of DAST, with the
formation of a UV-active baseline spot.
Product 14 appeared very early, but the
amount did not increase if the reaction time
was extended beyond a few hours. Reactions
in dichloromethane or toluene formed a pre-
cipitate within a few minutes of adding DAST
to 13. The yield of 14 was significantly lower
at 0 °C, while no reaction was observed below
this temperature. Elimination and decomposi-
tion were enhanced at temperatures above
ambient. Most significantly, if the reactions
were quenched with water, methanol or
aqueous sodium bicarbonate, the mass recov-
eries were very low, while quenching a reac-
tion conducted in tetrahydrofuran with 1 N
aqueous sodium hydroxide led to the reap-
pearance of the alcohol 13. This was our most
efficient reaction, giving 14 in 44% yield after
chromatography, along with 42% of recovered
13.
The high polarity and poor solubility of our
putative intermediate was unexpected, and it
seemed possible that the thioglycoside group
might be involved in some way. Ando et al.
reported essentially the same low yield from
DAST treatment of their thiofucoside [12],
whereas Lindhorst and Thiem obtained a 76%
yield of fluorine-containing products from a
very similar reaction of an O-benzoylated fu-
cose [18]. In order to see whether there was
some interaction between the reagent and the
thioglycoside in 13, we attempted to observe
the intermediate in tetrahydrofuran-d8 solu-
19
1
tion by F and H NMR spectrometry. Spec-
tra obtained at or below −30 °C showed only
signals due to DAST and 13, while at 0 °C or
above, product 14 was seen. No NMR signals
attributable to an intermediate could be de-
tected. The exact reasons for the sluggish reac-
tion of DAST with our fucoside remain
obscure.
This suggested to us that the intermediate
diethylaminosulfur difluoride ester that
formed during the reaction of 13 with DAST
was highly polar and unusually unreactive
either to displacement or hydrolysis. The base-
line TLC spot and the precipitates we ob-

18
P.G. Hultin, R.M. Buffie / Carbohydrate Research 322 (1999) 14–25
Since the polar intermediate did not de-
grade during the reaction and could be hy-
drolyzed back to 13 by 1 N aqueous sodium
hydroxide, we recycled the reaction to obtain
a high overall conversion of 13 to 14. Thus,
reaction with 1.5 equivalents of DAST at
room temperature in tetrahydrofuran gave
about 40% conversion after 2 h, as judged by
1H NMR analysis of the crude product. We
carried out this treatment sequentially three
more times, maintaining the concentrations of
unreacted substrate and of reagent essentially
constant at each cycle. This process, while
admittedly unwieldy, gave an 80% yield of 14
(based on the amount of 13 submitted to the
first reaction cycle) as an oil contaminated
with only a small amount of the more polar
elimination product 24. This oily product was
pure enough to be used directly in the next
step, but it could be crystallized to obtain pure
material if so desired.
The removal of the 2-O-benzoyl group from
14 under typical Zemple´n conditions was re-
markably sluggish, requiring 90 h and 3 equiv-
alents of sodium methoxide to proceed to
completion at room temperature. A molecular
mechanics-derived structure suggested that the
flanking equatorial 3-O-benzyl and 1-S-ethyl
groups of 14 were hindering access to the
2-position ester group. However, in boiling
methanolic sodium methoxide, the reaction
proceeded smoothly in only 1.5 h to form 15
(97%).
The 2-ulose product 16 is evidently very
prone to elimination of the 4-fluoride, despite
the fact that this process is formally an axial/
equatorial 1,2-syn elimination. Perlin and co-
workers noted this type of elimination many
years ago, arising from dimethyl sulfoxide–
SO3·pyridine oxidations of
D-glucosides [28].
We found that Swern oxidation of 15 was a
good method to obtain the 2-ulose 16, al-
though the conditions for the base-promoted
step of the reaction had to be carefully con-
trolled. The usual procedure of employing an
excess of triethylamine led to a complex
product mixture that included 26 as well as
ring-opened products. When a stoichiometric
amount of triethylamine was added gradually
to the reaction mixture at −78 °C, followed
by warming to room temperature over 45 min,
16 was obtained accompanied by little or no
elimination or degradation. NMR analysis of
the crude products obtained from several rep-
etitions of this reaction typically revealed
about 7% of unreacted 15. Chromatographic
purification of 16 was impossible because it
streaked on silica gel, presumably due to equi-
libration between the ketone and its hydrate.
The ketone 16 was immediately reduced
with sodium borohydride in 1:1 2-propanol–
dichloromethane at 0 °C, affording a 55:18
mixture of 17 and 15, from which pure 17 was
obtained by careful chromatography (54%;
63% correcting for recovered 15). Allowing for
the 7% of 15 present in the ketone, the selec-
tivity of this reduction was about 5:1. This
was disappointing, because Lichtenthaler and
Our approach to a 4-fluoro-
L
-rhamnoside
from an -fucose derivative required us to
L
invert the stereochemistry of the 2-position of
15. In principle, this could be done by an SN2
process, but Baer et al. observed that
analogous 2-triflates undergo ring contraction
when subjected to even mild nucleophilic dis-
placement conditions [24]. We therefore used
an oxidation–reduction sequence. The steric
constraints around the 2-OH group caused
some difficulties here also. Neither pyridinium
Schneider-Adams reported 20–50:1
D
-
manno: -gluco selectivities in several
D
analogous sodium borohydride reductions of
2-uloses [29]. Nevertheless, the reaction pro-
vided a reasonable yield of the desired
L
-
rhamno configuration, and the isomer could be
resubmitted to the oxidation–reduction proce-
dure. Benzoylation of the 2-OH group of 17
,
dichromate–acetic anhydride–3 A molecular
sieves [25,26] nor tetrapropylammonium per-
ruthenate–N-methylmorpholine N-oxide [27]
oxidized 15, even under forcing conditions.
Treatment of 15 with acetic anhydride and
dimethyl sulfoxide afforded predominantly the
a,b-unsaturated ketone 26, as did Pfitzner–
Moffatt conditions.
gave the desired 4-deoxy-4-fluoro- -rhamnosyl
L
donor 18 (99%).
Glycosidations of methyl 2-acetamido-4,6-
O-benzylidene-2-deoxy-a-
D
-glucopyranoside
[30] with donors 5 and 18 were performed in

P.G. Hultin, R.M. Buffie / Carbohydrate Research 322 (1999) 14–25
19
dichloromethane in the presence of N-iodo-
succinimide and triflic acid to afford the a-
linked disaccharides 27 (82%) and 28 (55%) as
single anomers. The disarming effect of the
4-fluoride in 18 was clearly evident in both the
yields and the rates of these reactions; the
reaction of 5 was complete in 15 min, whereas
that of 18 required 50 min to consume the
donor. The anomeric configurations of the
newly formed glycosidic bonds were assigned
materials and their biological activities will be
reported in future publications. We also antic-
ipate that the new glycosyl donors 5 and 18
will be useful to other workers exploring in-
hibitors of the synthesis of various bacterial
polysaccharides containing linkages to the 4-
position of -rhamnose.
L
3. Experimental
1
as a on the basis of the JC,H of approximately
171 Hz [31] observed in both 27 and 28.
General methods.—Reagents were obtained
from Aldrich Chemical Co. and were used as
received except as noted. Benzyl bromide was
passed through a plug of alumina immediately
before use. Dimethyl sulfoxide and N,N-
dimethylformamide (DMF) were dried over 3
,
A molecular sieves [32]. Triethylamine and
pyridine were dried by distillation from CaH2.
Deprotection of the disaccharides required
two steps in each case, and we found that the
order in which the reactions were performed
was crucial. Thus, treatment of 27 with 80%
acetic acid at 70 °C for 1.5 h cleanly hy-
drolyzed the benzylidene acetal. Subsequent
methoxide-promoted removal of the benzoate
groups afforded the unprotected disaccharide
29 in 95% yield. On the other hand, if the
benzoates were removed first, subsequent
acid-catalyzed hydrolysis of the acetal was
extremely sluggish, and after 8 h of heating
several byproducts were evident by thin-layer
chromatography (TLC) analysis. To deprotect
28, it was necessary first to remove the 2%-ben-
zoate, and subsequently to cleave both the
acetal and the benzyl ether by atmospheric
pressure hydrogenolysis in methanol over 10%
palladium-on-charcoal. The hydrogenolysis of
the benzyl ether occurred within 3 h, but
removal of the benzylidene acetal was rather
slow, requiring a further 20 h. Nevertheless,
an 81% yield of 30 was easily obtained after
chromatography. In contrast, we could not
induce hydrogenolysis of 28 to occur, even
under a hydrogen pressure of 3 atm.
N-Iodosuccinimide was recrystallized from
dioxane–CCl4.
Tetrahydrofuran
(THF),
CH2Cl2 and MeOH were dried according to
literature methods [33] before use; other sol-
vents were used as received. All reactions re-
quiring dry conditions were performed under
a positive pressure of dry nitrogen or argon, in
glassware dried at 120 °C for at least 18 h.
Organic layers from extractions were dried
using MgSO4 before concentration. ‘Chro-
matography’ refers to the use of the flash
procedure on Silica Gel 60 (230–400 mesh E.
Merck).
NMR spectra were acquired on Bruker AM
300 or AMX 500 instruments at 300 K (except
as noted) in the deuterated solvents indicated,
1
at 300.133 or 500.138 MHz for H, at 75.469
13
or 125.773 MHz for C, and at 282.467 MHz
1
for 19F. For H or 13C spectra, the residual
signals of the solvents were used as internal
chemical shift standards, except for spectra
acquired in D2O. In these cases, external refer-
19
encing to TSP was employed. The F spectra
were externally referenced to C6F6 and are
reported in ppm relative to CFCl3 at 0.00
ppm. All signal assignments given were confi-
rmed by analysis of J-couplings, COSY,
DEPT and/or 13C–1H HSQC experiments. In
ambiguous cases, coupling constants and
chemical shifts were verified by iterative simu-
The disaccharides 29 and 30 are prototypes
of potential inhibitors of mycobacterial AG
biosynthesis. We are now extending the chem-
istry described in this paper to disaccharides
having aglycons that better mimic the phos-
pholipid moiety of 1. The syntheses of these
lations using the computer program XSIM
written by Dr K. Marat.
,

20
P.G. Hultin, R.M. Buffie / Carbohydrate Research 322 (1999) 14–25
applied to a plug of silica gel (approximately
50 g) in a fritted funnel. The plug was eluted
with 4:1 hexanes–EtOAc to yield the thiocar-
bonylimidazolide 4, 1.16 g (95%).
Positive-ion mass spectra were obtained us-
ing 70-eV electron-impact ionization (EIMS,
HRMS) on a VG ZAB 7070E instrument, or
by electrospray ionization (ESIMS) on a
Micromass Quattro LC instrument at a cone
voltage of 20 V. Optical rotations were mea-
Tributyltin hydride (1 mL, 3.71 mmol) was
dissolved in dry toluene (30 mL). The mixture
was boiled, and a solution of 4 in toluene (11
mL) was added dropwise over 30 min. Heat-
ing was continued for 1 h, after which time
the reaction was cooled and the solvent was
evaporated. The residue was dissolved in
MeCN (100 mL), and the solution was
washed with hexanes (3×25 mL). Evapora-
tion of the MeCN layer afforded a crude oil
that was purified by chromatography (9:1 hex-
anes–EtOAc) to yield 758 mg (86%) of oily 5:
[h]D −35.2° (c 0.50, CHCl3); 1H NMR
(CDCl3): l 1.34 (tr, 3 H, J 7.4 Hz), 1.34 (d, 3
H, J5,6 6.2 Hz, H-6), 1.97–2.15 (m, 2 H, H-4ax,
H-4eq), 2.60–2.80 (m, 2 H), 4.45 (m, 1 H,
H-5), 5.48 (d, 1 H, J1,2 1.7 Hz, H-1), 5.51–5.58
(m, 2 H, H-2, H-3), 7.33 (m, 2 H), 7.45–7.52
(m, 3 H), 7.61 (m, 1 H), 7.89 (m, 2 H), 8.12
(m, 2 H); 13C NMR (CDCl3): l 15.07, 21.17,
25.59, 34.30, 64.90, 68.40, 70.42, 82.87, 128.32,
128.51, 129.66, 129.84, 130.02, 133.06, 133.29,
165.43, 165.69; Anal. Calcd for C22H24O5S: C,
65.98; H, 6.04; S, 8.01. Found: C, 66.06; H,
6.25; S, 7.89.
sured at ambient temperature using
a
Rudolph Research Autopol III polarimeter in
the indicated solvents. Elemental analyses
were performed by Chemisar Laboratories,
Guelph, Ont.
Ethyl 2,3-di-O-benzoyl-1-thio-h- -rhamno-
L
pyranoside (3).—Following a procedure of
Ando et al. [12], rhamnoside 2 [11] (1.00 g, 4.8
mmol) was dissolved in dry pyridine (3.5 mL)
and CH2Cl2 (7 mL). The solution was cooled
to −40 °C, and a solution of benzoyl chloride
(1.22 mL, 10.5 mmol) in CH2Cl2 (3 mL) was
added dropwise over 20 min. The mixture was
stirred for 30 additional min, and then MeOH
(5 mL) was added. The cooling bath was then
removed. When the mixture reached room
temperature, the solvents were evaporated,
and the residue was redissolved in CH2Cl2 (75
mL). The solution was washed with 1.2 N
HCl (3×15 mL), water (3×15 mL) and brine
(15 mL). The dried organic layer was evapo-
rated, and the residual oil was chro-
matographed (4:1 hexanes–EtOAc) to afford
538 mg of the tribenzoate and 1.45 g (72.5%)
Ethyl 2-O-benzoyl-3,4-O-isopropylidene-1-
1
of 3, as a viscous oil: H NMR (CDCl3): l
thio-i- -fucopyranoside (12).—A solution of
L
1.34 (tr, 3 H, J 7.4 Hz), 1.46 (d, 3 H, J5,6 6.1
Hz, H-6), 2.61–2.80 (m, 2 H), 3.97 (dd, 1 H,
J4,5 9.4, J3,4 9.8 Hz, H-4), 4.26 (dq, 1 H, H-5),
5.41 (d, 1 H, J1,2 1.4 Hz, H-1), 5.48 (dd, 1 H,
J2,3 3.4 Hz, H-3), 5.69 (dd, 1 H, H-2), 7.33 (m,
2 H), 7.45–7.53 (m, 3 H), 7.61 (m, 1 H), 7.90
11 [20] (3.00 g, 8.97 mmol) in dry MeOH (30
mL) was treated with sodium (20 mg). After
stirring for 45 min, Amberlite IR-120 (H+)
was added. The neutralized mixture was then
filtered and evaporated, removing the last
traces of methanol using a vacuum pump.
The resulting syrup was dissolved in dry
DMF (15 mL). To this solution was added
2,2-dimethoxypropane (15 mL) and p-toluene-
sulfonic acid hydrate (50.5 mg, 0.265 mmol),
and the mixture was allowed to stir for 2.5 h.
Triethylamine (2.5 mL) was added, after
which the solvents were evaporated on the
vacuum pump.
13
(m, 2 H), 8.08 (m, 2 H); C NMR (CDCl3): l
14.94, 17.64, 25.55, 69.34, 72.31, 72.74, 73.53,
82.08 (1JC,H 166.7 Hz), 128.41, 128.57, 129.28,
129.67, 129.86, 133.42, 133.48, 165.46, 166.74;
Anal. Calcd for C22H24O6S: C, 63.44; H, 5.81.
Found: C, 63.41; H, 5.87.
Ethyl 2,3-di-O-benzoyl-4,6-dideoxy-1-thio-
h- -lyxo-hexopyranoside (5).—Rhamnoside 3
L
(967 mg, 2.32 mmol) and thiocarbonyldiimi-
dazole (1.30 g, 7.3 mmol) were boiled in dry
toluene (50 mL) for 18 h. After cooling, the
supernatant was decanted from an insoluble
oily material and evaporated. The residue was
redissolved in a small amount of toluene and
The residue was dissolved in dry pyridine
(7.5 mL), and the solution was cooled to 0 °C.
Benzoyl chloride (1.6 mL, 13.8 mmol) was
added dropwise, and the reaction was stirred
overnight, gradually warming to room tem-

P.G. Hultin, R.M. Buffie / Carbohydrate Research 322 (1999) 14–25
21
7.18 (m, 5 H), 7.46 (m, 2 H), 7.59 (m 1 H),
8.02 (m 2 H); 13C NMR (CDCl3): l 14.82,
16.82, 23.51, 69.13, 69.51, 71.40, 74.62, 79.66,
83.11, 127.85, 127.99, 128.38, 128.46, 129.88,
130.05, 133.10, 137.21, 165.49; Anal. Calcd for
C22H26O5S: C, 65.65; H, 6.51; S, 7.97. Found:
C, 65.32; H, 6.75; S, 7.71.
perature. An excess of MeOH was added,
after which the solvents were evaporated. The
residue was dissolved in ether (100 mL), and
the solution was washed with 25-mL portions
of cold water, cold 1.2 N HCl, and satd aq
NaHCO3. The ether layer was dried and evap-
orated. The residue was purified by chro-
matography (6:1 hexanes–EtOAc) to yield 12,
2.71 g, (85.7%) as white crystals: mp 94–
Ethyl 2-O-benzoyl-3-O-benzyl-4,6-dideoxy-
4-fluoro-1-thio-i- -glucopyranoside (14).—
L
1
The reaction of 13 (2.00 g, 4.97 mmol) with
DAST in dry THF was performed in four
cycles. In each cycle, the substrate was dis-
solved in THF and stirred, while DAST was
added dropwise at rt. Stirring was continued
for 2 h. The reaction mixture was then trans-
ferred dropwise into a mixture of 1 N aq
NaOH (50 mL) and ether (100 mL), and
vigorously stirred for 10–15 min, after which
the phases were separated. The ether layer was
washed with water (2×25 mL), 1.2 N HCl
(25 mL), and satd aq NaHCO3 (25 mL). The
organic extract was then dried and evapo-
rated, and the residual oil was dried in vacuo.
95 °C; [h]D −48.1° (c 1.05, CHCl3); H NMR
(CDCl3): l 1.23 (tr, 3 H, J 7.4 Hz), 1.36 (s, 3
H), 1.45 (d, 3 H, J5,6 6.5 Hz, H-6), 1.62 (s, 3
H), 2.59–2.84 (m, 2 H), 3.94 (dq, 1 H, J4,5 2.0
Hz, H-5), 4.12 (dd, 1 H, J3,4 5.3 Hz, H-4), 4.30
(dd, 1 H, J2,3 7.3 Hz, H-3), 4.48 (d, 1 H, J1,2
10.1 Hz, H-1), 5.26 (dd, 1 H, H-2), 7.42 (m, 2
H), 7.54 (m, 1 H), 8.05 (m 2 H); 13C NMR
(CDCl3): l 14.74, 16.85, 23.83, 26.37, 27.86,
72.09, 72.74, 76.49, 77.29, 82.36, 110.09,
128.21, 129.79, 129.87, 132.97, 165.39; Anal.
Calcd for C18H24O5S: C, 61.34; H, 6.86; S,
9.10. Found: C, 61.61; H, 7.17; S, 9.08.
Ethyl 2-O-benzoyl-3-O-benzyl-1-thio-i- -
L
1
The extent of conversion was estimated by H
fucopyranoside (13).—Compound 12 (1.50 g,
4.26 mmol) was stirred for 3.5 h in 80% AcOH
(25 mL) at 50 °C. The solvent was evaporated,
after which dry toluene (3×20 mL) was
added and evaporated from the residue.
Dibutyltin oxide (1.17 g, 4.70 mmol) and dry
toluene (50 mL) were then added, and the
mixture was boiled for 2 h, with azeotropic
removal of water in a Dean–Stark trap. The
cooled solution was then evaporated.
NMR spectroscopy. The amounts of reagent
and solvent used were calculated to provide
roughly constant reactant concentrations, as-
suming 40% conversion at each cycle. These
are summarized in Table 1.
After the fourth cycle, the crude oil was
chromatographed (gradient elution from
5%“30% EtOAc–hexanes) to provide 14
(1.62 g, 80%) as a pale yellow oil contami-
nated by a small amount of elimination
product 24. After standing for several days,
the oil solidified. A pure sample of 14 recrys-
tallized as long needles from petroleum ether
(bp 30–60 °C): mp 59–60 °C; [h]D −14.1° (c
The residue was dissolved in dry DMF (15
mL). Anhydrous CsF (1.28 g, 8.49 mmol) was
added, followed by benzyl bromide (0.63 mL,
5.3 mmol), and the mixture was stirred over-
night at 40 °C. The solvents were evaporated,
and the residue was dissolved in ether (100
mL). The solution was washed with water
(3×20 mL), dried, and evaporated. Chro-
matography (3:1 hexanes–EtOAc) afforded 13
(1.63 g, 95%) as a thick oil that crystallized on
scratching and standing: mp 97–98 °C; [h]D
1
1.04, CHCl3); H NMR (CDCl3): l 1.23 (tr, 3
H, J 7.5 Hz), 1.42 (dd, 3 H, J6,F 1.3, J5,6 6.2
Hz, H-6), 2.62–2.79 (m, 2 H), 3.59 (ddd, 1 H,
J5,F 3.0, J4,5 9.1 Hz, H-5), 3.83 (ddd, 1 H, J3,F
Table 1
Reaction of ethyl 2-O-benzoyl-3-O-benzyl-1-thio-b-
ranoside (13) with DAST
L-fucopy-
1
−13.6° (c 1.03, CHCl3); H NMR (CDCl3): l
1.20 (tr, 3 H, J 7.4 Hz), 1.41 (d, 3 H, J5,6 6.5
Hz, H-6), 2.52 (br s, 1 H, OH), 2.64–2.83 (m,
2 H), 3.62–3.69 (m, 2 H, H-3, H-5), 3.91 (br s,
1 H, H-4), 4.46 (d, 1 H, J1,2 10.0 Hz, H-1),
4.55 (d, 1 H, J 12.3 Hz), 4.67 (d, 1 H, J 12.3
Hz), 5.50 (dd, 1 H, J2,3 9.3 Hz, H-2), 7.16–
Cycle
THF (mL)
DAST [mL (mmol)]
1
2
3
4
20
12
7
990 (7.5)
600 (4.5)
360 (2.7)
220 (1.7)
5

22
P.G. Hultin, R.M. Buffie / Carbohydrate Research 322 (1999) 14–25
was stirred for 15 min, and then a solution of
15 (600 mg, 2.00 mmol) in CH2Cl2 (2.5 mL+
a 2.5 mL rinse) was added dropwise over 10
min. After stirring the reaction at −78 °C for
a further 30 min, Et3N (615 mL, 4.40 mmol)
was added over 3 min. The cloudy mixture
was kept at −78 °C for 15 min, and then the
cold bath was removed. The reaction mixture
reached rt over 45 min, after which time it was
diluted with ether (50 mL) and washed with
water (2×10 mL) and brine (10 mL). The
dried solution was evaporated to afford a
yellow oily residue (608 mg). The product was
identified by NMR spectroscopy as a mixture
of unreacted 15 (approximately 7%) and ethyl
14.4, J3,4 8.9, J2,3 9.2 Hz, H-3), 4.29 (ddd, 1 H,
J4,F 50.1 Hz, H-4), 4.53 (d, 1 H, J1,2 9.8 Hz,
H-1), 4.67 (d, 1 H, J 11.5 Hz), 4.78 (d 1 H, J
11.5 Hz), 5.26 (dd, 1 H, H-2), 7.12–7.19 (m, 5
H), 7.46 (m, 2 H), 7.60 (m, 1 H), 8.01 (m, 2
13
H); C NMR (CDCl3): l 14.76, 17.53, 23.91,
71.58 (d, JC,F 9.4 Hz), 73.90 (d, JC,F 20.0 Hz),
74.07, 80.51 (d, JC,F 17.7 Hz), 83.24, 95.15 (d,
JC,F 184.3 Hz), 127.62, 127.95, 128.17, 128.26,
128.33, 128.41, 129.70, 129.84, 133.15, 137.48,
19
165.04; F NMR (CDCl3): l −197.10 (dddq,
J6,F 1.3, J5,F 3.0, J3,F 14.4, J4,F 50.1 Hz); EIMS:
m/z 343 [M–SEt]+, 235, 176, 149, 105, 91, 77;
HRMS on m/z 343 fragment: Calcd for
[C20H20FO4]+: 343.1345. Found: 343.1347;
Anal. Calcd for C22H25FO4S: C, 65.33; H,
6.23; F, 4.70. Found: C, 64.97; H, 6.53; F,
4.89.
3-O-benzyl-4-deoxy-4-fluoro-1-thio-b- -ara-
L
bino-hex-2-ulopyranoside (16): 1H NMR
(CDCl3): l 1.30 (tr, 3 H, J 7.5 Hz), 1.46 (dd,
3 H, J6,F 1.4, J5,6 6.2 Hz, H-6), 2.71 (app q, 2
H, Japp 7.5 Hz), 3.83 (ddq, 1 H, J5,F 2.8, J4,5
9.2 Hz, H-5), 4.17 (dd, 1 H, J3,4 8.9, J3,F 17.0
Hz, H-3), 4.36 (ddd, 1 H, J4,F 50.5 Hz, H-4),
4.71 (d, 1 H, J 11.8 Hz), 4.96 (d, 1 H, J 11.8
Hz), 5.04 (d, 1 H, J1,F 0.4 Hz, H-1), 7.30–7.43
(m, 5 H); 13C NMR (CDCl3); l 14.75, 17.57,
23.80, 73.35, 74.21 (d, JC,F 23.9 Hz), 83.44 (d,
JC,F 18.5 Hz), 86.19, 95.04 (d, JC,F 192.4 Hz),
127.91, 128.00, 128.40, 136.85, 196.25 (d, JC,F
8.0 Hz).
Ethyl 3-O-benzyl-4,6-dideoxy-4-fluoro-1-
thio-i- -glucopyranoside (15).—A solution of
L
14 (550 mg, 1.36 mmol) in dry MeOH (25 mL)
was treated with sodium (100 mg, 4.35 mmol).
The reaction was gently boiled under reflux
for 1.5 h. Amberlite IR-120 (H+) was added
to the cooled solution to neutralize the base,
after which the mixture was filtered and evap-
orated to dryness. The resulting yellow oil was
chromatographed (6:1 hexanes–EtOAc) to af-
ford 397 mg (97%) of crystalline 15: mp:
1
The crude ketone 16 was dissolved in 1:1
CH2Cl2–2-PrOH (5 mL), and the solution was
cooled in an ice bath. Finely powdered
NaBH4 (115 mg, 3.05 mmol) was added grad-
ually, and the mixture was stirred while warm-
ing slowly to rt. After 4.5 h, TLC analysis
indicated complete conversion to a mixture of
two products. The reaction mixture was
cooled on ice and quenched by dropwise addi-
tion of 15% AcOH until all effervescence
ceased. Ether (50 mL) was then added, and
the solution was washed with water (5 mL)
and satd aq NaHCO3 (5 mL). The dried solu-
tion was evaporated to provide a clear oil.
High-performance liquid chromatography
(HPLC) analysis (Waters m-Porasil 3.9×300
mm column, linear gradient elution from 0.1
35.5–37 °C; [h]D +43.9° (c 0.82 CHCl3); H
NMR (CDCl3): l 1.31 (tr, 3 H, J 7.4 Hz), 1.37
(dd, 3 H, J6,F 1.4, J5,6 6.1 Hz, H-6), 2.44 (m, 1
H, OH), 2.65–2.81 (m, 2 H), 3.47–3.63 (m, 3
H, H-2, H-3, H-5), 4.16 (ddd, 1 H, J 8.5, J
9.4, J4,F 50.1 Hz, H-4), 4.34 (d, 1 H, J1,2 9.5
Hz, H-1), 4.81 (d, 1 H, J 11.5 Hz), 4.90 (d, 1
13
H, J 11.5 Hz), 7.30–7.42 (m, 5 H); C NMR
(CDCl3): l 15.14, 17.49, 24.29, 72.52 (d, JC,F
9.1 Hz), 73.76 (d, JC,F 23.7 Hz), 74.42 (d, JC,F
1.9 Hz), 82.81 (d, JC,F 17.2 Hz), 85.45, 94.55
(d, JC,F 183.9 Hz), 127.74, 127.89, 128.34,
138.02; 19F NMR (CDCl3): l −193.54 (app
dd, Japp 13.3, Japp 50.2 Hz); Anal. Calcd for
C15H21FO3S: C, 59.98; H, 7.05. Found: C,
59.66; H, 7.17.
Ethyl 3-O-benzyl-4-deoxy-4-fluoro-1-thio-i-
-rhamnopyranoside (17).—Oxalyl chloride
to 0.5% 2-PrOH–hexanes at 1.00 mL min−1
,
L
UV detection at 254 nm) showed a 55:18 ratio
of major products, along with several byprod-
ucts. Careful flash chromatography of the
(225 mL, 2.58 mmol) was added dropwise to a
cold (−78 °C) solution of Me2SO (185 mL,
2.60 mmol) in CH2Cl2 (5 mL). The mixture

P.G. Hultin, R.M. Buffie / Carbohydrate Research 322 (1999) 14–25
23
Methyl 2,3-di-O-benzoyl-4,6-dideoxy-h-
lyxo - hexopyranosyl - (1“3) - 2 - acetamido - 4,
6-O-benzylidene-2-deoxy-h- -glucopyranoside
L
-
mixture (gradient elution, 10“20% EtOAc–
hexanes) afforded 15 (79 mg, 13%) and 17
(326 mg, 54%), which crystallized on standing:
mp: 78–79.2 °C; [h]D +57.3° (c 1.03, CHCl3);
1H NMR (CDCl3): l 1.30 (tr, 3 H, J 7.5 Hz),
1.39 (dd, 3 H, J6,F 1.0, J5,6 6.1 Hz, H-6), 2.57
(br s, OH), 2.73 (app q, 2 H, Japp 7.4 Hz), 3.44
(ddq, 1 H, J5,F 2.5, J4,5 9.2 Hz, H-5), 3.57
(ddd, 1 H, J2,3 3.0, J3,4 9.0, J3,F 13.9 Hz, H-3),
4.11 (m, 1 H, H-2), 4.46 (ddd, 1 H, J4,F 50.7
Hz, H-4), 4.55 (br s, 1 H, H-1), 4.70 (d, 1 H,
J 11.9 Hz), 4.80 (d, 1 H, J 11.9 Hz), 7.31–7.37
(m, 5 H); 13C NMR (CDCl3): l 14.96, 17.45,
25.47, 71.09 (d, JC,F 9.4 Hz), 72.18, 74.03
(d, JC,F 24.5 Hz), 79.06 (d, JC,F 15.1 Hz),
83.25, 93.02 (d, JC,F 180.7 Hz), 127.80, 128.02,
128.50, 137.35; Anal. Calcd for C15H21-
FO3S: C, 59.98; H, 7.05. Found: C, 59.70; H,
7.20.
D
(27).—Thioglycoside 5 (200 mg, 0.50 mmol),
methyl 2-acetamido-4,6-O-benzylidene-2-de-
oxy-a- -glucopyranoside [30] (194 mg, 0.60
D
,
mmol) and powdered, freshly activated 4 A
molecular sieves (500 mg) were placed in a dry
two-necked flask fitted with a septum, a three-
way gas inlet valve and a magnetic stirring bar
and left overnight at 0.1 torr. The flask was
then flushed with argon, dry CH2Cl2 (4.0 mL)
was added, and the mixture was vigorously
stirred for 1 h. Recrystallized N-iodosuccin-
imide (146 mg, 0.65 mmol) was added, fol-
lowed by TfOH (5.8 mL, 0.065 mmol). TLC
analysis indicated complete consumption of
the donor after 15 min. The mixture was
filtered through Celite into 50% satd aq
Na2S2O3 (5 mL) in a separatory funnel. The
filter cake was washed with several portions of
CH2Cl2 (45 mL total), which were also added
to the separatory funnel. The mixture was
shaken until all color was discharged, and the
phases were separated. The organic layer was
washed with satd aq NaHCO3 (5 mL), water
(5 mL) and brine (5 mL), then dried and
concentrated. Chromatography of the solid
residue (3:1 CH2Cl2 –EtOAc) afforded 27 (272
mg, 82%) as white crystals: mp 230–231 °C;
[h]D +32.0° (c 0.5, CHCl3); 1H NMR
(CDCl3): l 0.76 (d, 3 H, J5%,6% 6.2 Hz, H-6%),
1.78–1.98 (m, 2 H, H-4%, H-4a%), 2.17 (s, 3 H)
3.40 (s, 3 H), 3.69 (dd, 1 H, J4,5 8.6, J3,4 9.8
Hz, H-4), 3.76–3.90 (m, 2 H, H-5%, H-6), 3.95
(dd, 1 H, J2,3 10.0 Hz, H-3), 4.18–4.32 (m, 2
H, H-5, H-6a), 4.47 (ddd, 1 H, J1,2 3.7, JNH,2
10.0 Hz, H-2), 4.66 (d, 1 H, H-1), 5.17–5.20
(m, 2 H, H-1%, H-2%), 5.54 (ddd, 1 H, J2%,3% 3.0,
J4%,3% 5.1, J4a%,3% 11.8 Hz, H-3%), 5.57 (s, 1 H),
5.77 (d, 1 H, NH), 7.30–7.37 (m, 5 H), 7.43–
7.53 (m, 5 H), 7.59 (m, 1 H), 7.89 (m, 2 H),
8.07 (m, 2 H); 13C NMR (CDCl3): l 20.35,
23.51, 33.71, 53.15, 55.17, 63.11, 64.25, 67.76,
68.97, 69.40, 76.06, 80.32, 99.08 (1JC,H 171.6
Hz), 99.43 (1JC,H 170.2 Hz), 102.15, 126.42,
128.13, 128.21, 128.40, 129.09, 129.56, 129.67,
129.94, 130.00, 132.91, 133.14, 137.22, 165.40,
165.62, 170.57; Anal. Calcd for C36H39NO11:
C, 65.35; H, 5.94; N, 2.12. Found: C, 65.68;
H, 6.21, N, 2.09.
Ethyl 2-O-benzoyl-3-O-benzyl-4-deoxy-4-
fluoro-1-thio-i- -rhamnopyranoside (18).—A
L
mixture of 17 (200 mg, 0.67 mmol), dry
pyridine (2.5 mL) and DMAP (8 mg, 0.07
mmol) was treated with benzoyl chloride (230
mL, 1.98 mmol). The reaction was stirred for
2.5 h at rt. It was then cooled on a water bath,
while cold water (2.5 mL) was added drop-
wise. The mixture was stirred for 1 h more,
and then diluted with ether (30 mL). The
phases were separated, and the organic layer
was washed with 1.2 N HCl (4×5 mL) and
satd aq Na2CO3 (5 mL). Drying and evapora-
tion (finishing on a vacuum pump to remove
all traces of solvent) provided 270 mg (99%)
of 18 as a pure oil: [h]D +141.1° (c 1.00,
1
CHCl3); H NMR (CDCl3): l 1.28 (tr, 3 H, J
7.4 Hz), 1.48 (dd, 3 H, J6,F 1.3, J5,6 6.1 Hz,
H-6), 2.65–2.82 (m, 2 H), 3.56 (ddq, 1 H, J5,F
3.2, J4,5 9.2 Hz, H-5), 3.75 (ddd, 1 H, J2,3 3.7,
J3,4 9.2, J3,F 12.9 Hz, H-3), 4.47 (ddd, 1 H, J4,F
50.5 Hz, H-4), 4.62 (d, 1 H, J 12.2 Hz), 4.76
(d, 1 H, J1,2 1.0 Hz, H-1), 4.81 (d, 1 H, J 12.2
Hz), 5.88 (ddd, 1 H, J2,F 2.7 Hz, H-2), 7.28–
7.33 (m, 5 H), 7.45 (m, 2 H), 7.57 (m, 1 H),
8.10 (m, 2 H); 13C NMR (CDCl3): l 14.83,
17.79, 25.64, 71.20 (d, JC,F 9.2 Hz), 71.47,
74.40 (d, JC,F 24.0 Hz), 77.72 (d, JC,F 18.2 Hz),
82.40, 92.29 (d, JC,F 182.1 Hz), 127.74, 128.33,
129.48, 130.03, 133.18, 137.38, 165.72; Anal.
Calcd for C22H25FO4S: C, 65.33; H, 6.23.
Found: C, 65.60; H, 6.67.

24
P.G. Hultin, R.M. Buffie / Carbohydrate Research 322 (1999) 14–25
Methyl 2-O-benzoyl-3-O-benzyl-4-deoxy-4-
fluoro-h- -rhamnopyranosyl-(1“3)-2-aceta-
mido-4,6-O-benzylidene-2-deoxy-h- -glucopy-
and then dried on the vacuum pump.
L
Methanol (5 mL) was added, followed by
sodium (approximately 4 mg, 0.17 mmol). The
reaction was complete after 4 h of stirring at
rt. Amberlite IR-120 (H+) was added to neu-
tralize the base, and the mixture was filtered
and concentrated. Chromatography of the
residue (4:1 CHCl3–MeOH) afforded the un-
protected disaccharide 29 (78.8 mg, 95%) as a
white crystalline solid: mp 255–256 °C; [h]D
D
ranoside (28).—Using the procedure described
above for the synthesis of 27, the coupling of
fluororhamnose 18 (128 mg, 0.32 mmol) and
methyl 2-acetamido-4,6-O-benzylidene-2-de-
oxy-a- -glucopyranoside [30] (123 mg, 0.38
D
mmol) in CH2Cl2 (2.5 mL) was effected in the
,
presence of 4 A molecular sieve powder (350
1
mg), recrystallized N-iodosuccinimide (92 mg,
0.41 mmol) and TfOH (3.6 mL, 0.041 mmol).
TLC analysis indicated complete consumption
of the donor after 50 min. Workup and chro-
matography (3:2 EtOAc–hexanes) afforded
disaccharide 28 (116 mg, 55%) as an amor-
+11.2° (c 0.47, water); H NMR (310 K,
D2O): l 1.16 (d, 3 H, J5%,6% 6.2 Hz, H-6%), 1.56
(ddd, 1 H, J4%ax,4%eq 11.7, J4%ax,5% 11.7, J4%ax,3% 12.0
Hz, H-4%ax), 1.71 (m, 1 H, H-4%eq), 2.05 (s, 3
H, C(O)CH3), 3.40 (s, 3 H, OCH3), 3.51 (dd, 1
H, J3,4 8.9, J4,5 10.1 Hz, H-4), 3.60 (m, 1 H,
H-2%), 3.69 (ddd, 1 H, J5,6a 2.3, J5,6 5.4 Hz,
H-5), 3.72 (dd, 1 H, J2,3 10.4 Hz, H-3), 3.78
(dd, 1 H, J6,6a 12.2 Hz, H-6), 3.88 (dd, 1 H,
H-6a), 3.98 (ddd 1 H, J2%,3% 3.0, J3%,4%eq 4.8 Hz,
H-3%), 4.04 (dd, 1 H, J1,2 3.6 Hz, H-2), 4.21
(ddq, 1 H, J4%eq,5% 2.2 Hz, H-5%), 4.71 (d, 1 H,
H-1), 4.89 (d, 1 H, J1%,2% 1.9 Hz, H-1%); 13C
NMR (310 K, D2O): l 20.62 (C-6%), 22.66
(C(O)CH3), 35.15 (C-4%), 53.82 (C-2), 55.90
(OCH3), 61.38 (C-6), 65.85 (C-3%), 66.38 (C-
5%), 69.18 (C-2%), 69.32 (C-4), 72.64 (C-5),
80.15 (C-3), 99.01 (1JC,H 171.8 Hz, C-1),
102.76 (1JC,H 168.4 Hz, C-1%), 174.86 (CꢀO);
ESIMS m/z 388.25 [M+Na]+, 366.21 [M+
H]+, 236.00 [GlcNAcOCH3+H]+; Anal.
Calcd for C15H27NO9: C, 49.31; H, 7.45; N,
3.83. Found: C, 49.19; H, 7.60; N, 3.72.
1
phous solid: [h]D +53.0° (c 1.03, CHCl3); H
NMR (CDCl3): l 0.80 (dd, 3 H, J6%,F 1.2, J5%,6%
6.1 Hz, H-6%), 2.16 (s, 3 H), 3.40 (s, 3 H), 3.65
(dd, 1 H, J3,4 9.3, J4,5 9.3 Hz, H-4), 3.77 (dd, 1
H, J6,6a 9.9, J5,6 10.3 Hz, H-6), 3.81–3.84 (m,
1 H, H-5), 3.88 (dd, 1 H, J2,3 10.1 Hz, H-3),
4.02–4.08 (m, 2 H, H-3%, H-5%), 4.30 (dd, 1 H,
J5,6a 4.4 Hz, H-6a), 4.41 (ddd, 1 H, J3%,4% 9.5,
J4%,5% 9.5, J4%,F 51.4 Hz, H-4%), 4.45 (ddd, 1 H,
J1,2 3.8, JNH,2 10.0 Hz, H-2), 4.64 (d, 1 H, J
11.9 Hz), 4.65 (d, 1 H, H-1), 4.68 (d, 1 H, J
11.9 Hz), 4.98 (dd, 1 H, J1%,2% 2.0, J1%,F 2.8 Hz,
H-1%), 5.33 (ddd, 1 H, J2%,3% 3.2, J2%,F 3.2 Hz,
H-2%), 5.55 (s, 1 H), 5.72 (d, 1 H, NH),
13
7.19–7.57 (m, 13 H), 7.98 (m, 2 H); C NMR
(CDCl3): l 16.74 (C-6%), 23.59 (C(O)CH3),
53.14 (C-2), 55.26 (OCH3), 63.18 (C-5), 66.09
(d, JC,F 25.0 Hz, C-5%), 68.99 (C-6), 70.92 (d,
JC,F 9.1 Hz, C-2%), 71.74 (OCH2Ar), 74.99 (d,
JC,F 18.3 Hz, C-3%), 77.15 (C-3), 80.17 (C-4),
92.88 (d, JC,F 179.9 Hz, C-4%), 98.67 (1JC,H
171.0 Hz, C-1%), 99.49 (1JC,H 171.3 Hz, C-1),
102.13 (OCO), 126.40, 127.54, 127.63, 128.28,
128.30, 128.43, 129.25, 129.75, 129.83, 133.28,
137.13, 138.03, 165.67 (ester CꢀO), 170.63
(amide CꢀO); Anal. Calcd for C36H40FNO10:
C, 64.95; H, 6.06; N, 2.10. Found: C, 64.82;
H, 6.01; N, 2.01.
Methyl 4-deoxy-4-fluoro-h-
L
-rhamnopyra-
-glu-
nosyl-(1“3)-2-acetamido-2-deoxy-h-
D
copyranoside (30).—Protected disaccharide 28
(114 mg, 0.17 mmol) was dissolved in dry
MeOH (5 mL) and treated with sodium (2.5
mg, 0.11 mmol). The mixture was stirred for
16 h, then neutralized with Amberlite IR-120
(H+) and filtered into a fresh flask containing
a magnetic stirring bar. The resin was washed
with an additional 5 mL of MeOH, and the
washings were also added to the flask. Pd/C
catalyst (10%; 20 mg) was added to this solu-
tion (Caution! Extreme fire hazard!), the flask
was fitted with a three-way stopcock, and
evacuated on a water aspirator. A balloon
containing hydrogen was attached to the stop-
cock and the flask was charged with the gas.
The reaction was vigorously stirred for 23 h,
at which point TLC indicated complete re-
Methyl 4,6-dideoxy-h-
L
-lyxo-hexopyranosyl-
- gluco-
(1“3) - 2 - acetamido - 2 - deoxy - h -
D
pyranoside (29).—Protected disaccharide 27
(150 mg, 0.227 mmol) was suspended in 80%
AcOH (10 mL), and the mixture was heated at
70 °C for 1.5 h. The resulting clear solution
was then cooled and evaporated. The residue
was re-evaporated from toluene (3×5 mL)

P.G. Hultin, R.M. Buffie / Carbohydrate Research 322 (1999) 14–25
25
[2] G.S. Besra, P.J. Brennan, Biochem. Soc. Trans., 25
(1997) 845–850.
[3] P.J. Brennan, G.S. Besra, Biochem. Soc. Trans., 25
(1997) 188–194.
[4] M. McNeil, M. Daffe, P.J. Brennan, J. Biol. Chem., 265
(1990) 18200–18206.
[5] K. Mikusova, M. Mikus, G.S. Besra, I. Hancock, P.J.
Brennan, J. Biol. Chem., 271 (1996) 7820–7828.
[6] D.B. Young, K. Duncan, Annu. Re6. Microbiol., 49
(1995) 641–673.
[7] T. Nishio, Y. Miyake, K. Kubota, M. Yamai, S. Miki,
T. Ito, T. Oku, Carbohydr. Res., 280 (1996) 357–363.
[8] M. Bols, I. Lundt, E.R. Ottosen, Carbohydr. Res., 222
(1991) 141–149.
[9] V. Poszgay, A. Neszme´lyi, Carbohydr. Res., 85 (1980)
143–150.
[10] K. Kefurt, Z. Kefurtova´, J. Jary´, Collect. Czech. Chem.
Commun., 36 (1971) 1701–1704.
[11] F.-I. Auzanneau, D.R. Bundle, Carbohydr. Res., 212
(1991) 13–24.
[12] T. Ando, H. Ishida, M. Kiso, A. Hasegawa, Carbohydr.
Res., 249 (1993) 275–280.
[13] J. Defaye, A. Gadelle, S.J. Angyal, Carbohydr. Res., 126
(1984) 165–169.
[14] D.R. Bundle, M. Gerken, T. Peters, Carbohydr. Res., 174
(1988) 239–251.
[15] B. Doboszewski, G.W. Hay, W.A. Szarek, Can. J.
Chem., 65 (1987) 412–419.
[16] A.C. Richardson, Carbohydr. Res., 10 (1969) 395–402.
[17] P.A. Zunszain, O. Varela, Carbohydr. Res., 222 (1991)
131–140.
[18] T.K. Lindhorst, J. Thiem, Carbohydr. Res., 209 (1991)
119–129.
[19] P. Smid, G.A. de Ruiter, G.A. van der Marel, F.M.
Rombouts, J.H. van Boom, J. Carbohydr. Chem., 10
(1991) 833–849.
[20] G.H. Veeneman, S.H. van Leeuwen, J.H. van Boom, Te-
trahedron Lett., 31 (1990) 1331–1334.
moval of all protecting groups. The mixture
was filtered through Celite and evaporated.
Chromatography of the residue on a short
column (gradient elution from 15“30%
MeOH–CHCl3) afforded 30 (53.3 mg, 81%).
The product could be recrystallized from
MeCN: mp 253 °C (dec.); [h]D +16.0° (c 0.50,
1
water); H NMR (D2O, 310 K): l 1.27 (dd, 3
H, J6%,F 1.2, J5%,6% 6.2 Hz, H-6%), 2.05 (s, 3 H,
C(O)CH3), 3.40 (s, 3 H, OCH3), 3.54 (dd, 1 H,
J3,4 8.9, J4,5 10.1 Hz, H-4), 3.69 (ddd, 1 H, J5,6a
2.3, J5,6 5.4 Hz, H-5), 3.75 (dd, 1 H, J2,3 10.4
Hz, H-3), 3.78 (dd, 1 H, J6,6a 12.2 Hz, H-6),
3.82 (ddd, 1 H, J2%,F 1.8, J1%,2% 3.0, J2%,3% 3.5 Hz,
H-2%), 3.88 (dd, 1 H, H-6a), 3.99 (ddd, 1 H,
J3%,4% 9.3, J3%,F 14.5 Hz, H-3%), 4.05 (dd, 1 H, J1,2
3.6, J2,3 10.4 Hz, H-2), 4.18 (ddq, 1 H, J5%,F
4.8, J4%,5% 9.6 Hz, H-5%), 4.31 (ddd, 1 H, J4%,F
51.5 Hz, H-4%), 4.71 (d, 1 H, H-1), 4.85 (dd, 1
H, J1%,F 2.0 Hz, H-1%); 13C NMR (D2O): l
16.85 (C-6%), 22.68 (C(O)CH3), 53.88 (C-2),
55.92 (OCH3), 61.37 (C-6), 67.01 (d, JC,F 24.9
Hz, C-5%), 69.20 (C-4), 69.52 (d, JC,F 17.6 Hz,
C-3%), 72.11 (d, JC,F 9.3 Hz, C-2%), 72.72 (C-5),
80.36 (C-3), 94.14 (d, JC,F 175.0 Hz, C-4%),
99.00 (1JC,H 171.9 Hz, C-1), 101.81 (1JC,H 170.0
19
Hz, C-1%), 174.89 (CꢀO); F NMR (D2O, 300
K): l −204.18 (m); ESIMS m/z 406.17 [M+
Na]+, 384.26 [M+H]+, 236.18 [GlcNAcO-
CH3+H]+; Anal. Calcd for C15H26FNO9·
H2O: C, 44.89; H, 7.03; N, 3.49. Found: C,
44.83, H, 7.01; N, 3.39.
[21] K.C. Mange, W.J. Middleton, J. Fluor. Chem., 43 (1989)
405–413.
[22] T. Tsuchiya, Ad6. Carbohydr. Chem. Biochem., 48 (1990)
91–277.
[23] S.S. Yang, T.R. Beattie, T.Y. Shen, Tetrahedron Lett., 23
(1982) 5517–5520.
[24] H.H. Baer, F. Herna´ndez Mateo, L. Siemsen, Carbohydr.
Res., 187 (1989) 67–92.
Acknowledgements
[25] F. Andersson, B. Samuelsson, Carbohydr. Res., 129
(1984) C1–C3.
This research was supported by a Research
Grant from the Natural Sciences and Engi-
neering Research Council of Canada (to
P.G.H.) and a Manitoba CareerStart Grant
(to R.M.B). We thank Dr Kirk Marat and Mr
Terry Wolowiec for assistance in obtaining
NMR spectra. We also acknowledge Dr He´-
le`ne Perreault and Mr Wayne Buchannon for
providing the mass spectra used in this work.
[26] S. Czernecki, C. Georgoulis, C.L. Stevens, K. Vijayku-
marian, Tetrahedron Lett., 26 (1985) 1699–1702.
[27] W.P. Griffith, S.V. Ley, G.P. Whitcombe, A.D. White, J.
Chem. Soc., Chem. Commun., (1987) 1625–1627.
[28] G.M. Cree, D.M. Mackie, A.S. Perlin, Can. J. Chem., 47
(1969) 511–512.
[29] F.W. Lichtenthaler, T. Schneider-Adams, J. Org. Chem.,
59 (1994) 6728–6734.
[30] K. Yamamoto, T. Hayashi, Bull. Chem. Soc. Jpn., 46
(1973) 656–657.
[31] K. Bock, C. Pedersen, J. Chem. Soc., Perkin Trans. 2,
(1974) 293–297.
[32] D.R. Burfield, R.H. Smithers, J. Org. Chem., 43 (1978)
3966–3968.
References
[1] P.J. Brennan, H. Nikaido, Annu. Re6. Biochem., 64
[33] D.D. Perrin, W.L.F. Armarego, Purification of Labora-
tory Chemicals, third ed., Pergamon, Oxford, 1988.
(1995) 29–63.