2,3-Anhydrosugars in glycoside bond synthesis: Mechanism of 2-deoxy-2-thioaryl glycoside formation

Article abstract of DOI:10.1021/ja9029945

A series of investigations probing the mechanism of the 2,3-anhydrosugar migration - glycosylation reaction were performed using a thioglycoside with the D-lyxo stereochemistry as the substrate. Among the work reported are the results of quantum mechanical calculations, NMR studies, the measurement of α-deuterium kinetic isotope effects, and the synthesis of a series of substrate analogues. All studies point to a consistent finding: that the reaction proceeds through an oxocarbenium ion intermediate, not an episulfonium ion as previously suggested. It is proposed that the high stereoselectivity of the reaction arises from a preferred "inside attack" of the nucleophile onto the oxocarbenium ion intermediate.

Full text of DOI:10.1021/ja9029945

Published on Web 08/24/2009
2,3-Anhydrosugars in Glycoside Bond Synthesis: Mechanism
of 2-Deoxy-2-thioaryl Glycoside Formation
Dianjie Hou, Hashem A. Taha, and Todd L. Lowary*
Alberta Ingenuity Centre for Carbohydrate Science and Department of Chemistry, The UniVersity of
Alberta, Gunning-Lemieux Chemistry Centre, Edmonton, Alberta, T6G 2G2 Canada
Received April 15, 2009; E-mail: tlowary@ualberta.ca
Abstract: A series of investigations probing the mechanism of the 2,3-anhydrosugar migration-glycosylation
reaction were performed using a thioglycoside with the D-lyxo stereochemistry as the substrate. Among
the work reported are the results of quantum mechanical calculations, NMR studies, the measurement of
R-deuterium kinetic isotope effects, and the synthesis of a series of substrate analogues. All studies point
to a consistent finding: that the reaction proceeds through an oxocarbenium ion intermediate, not an
episulfonium ion as previously suggested. It is proposed that the high stereoselectivity of the reaction arises
from a preferred “inside attack” of the nucleophile onto the oxocarbenium ion intermediate.
Chart 1
Introduction
Over the past several years we have published a series of
papers demonstrating the utility of 2,3-anhydrosugar derivatives
(e.g., 1-7, Chart 1) in the synthesis of glycosidic bonds.1-6
One reaction manifold available to these species is the assembly
of 2,3-anhydosugar glycosides (e.g., 7). Postglycosylation
nucleophilic opening of the epoxide ring can lead to a diverse
range of products including R-galactofuranosides,4 ꢀ-arabino-
furanosides,5,6 and R-arabinofuranosides.1,6 Another pathway,
possible only for the 2,3-anhydrosugar thioglycosides (e.g., 1),
is the formation of 2-deoxy-2-thioaryl glycosides (9), which are
useful precursors to 2-deoxy-glycosides (10).2,3
The formation of 2-deoxy-2-thioaryl glycosides occurs when
thioglycosides such as 1, 2, and 4-6 are treated with an alcohol
and a Lewis acid. The major, usually exclusive, product is the
one in which the glycosidic bond is formed trans to the thioaryl
group at C-2 (e.g., 9). On the basis of this observation, and by
analogy to other similar processes,7-16 we proposed that the
mechanism of this migration-glycosylation reaction proceeds
via the pathway shown in Figure 1, pathway A.2 Activation of
the epoxide in 1 by the Lewis acid to generate 11 initiates the
migration of the thioaryl group from C-1 to C-2 via an
episulfonium ion intermediate 12 that, in turn, reacts with an
alcohol to generate the glycoside 9 with a high degree of
stereocontrol.
(1) Cociorva, O. M.; Lowary, T. L. Tetrahedron 2004, 60, 1481–1489.
(2) Hou, D.; Lowary, T. L. Org. Lett. 2007, 9, 4487–4490.
(3) Hou, D.; Lowary, T. L. J. Org. Chem. 2009, 74, 2278–2289.
(4) Bai, Y.; Lowary, T. L. J. Org. Chem. 2006, 71, 9658–9671.
(5) Callam, C. S.; Gadikota, R. R.; Krein, D. M.; Lowary, T. L. J. Am.
Chem. Soc. 2003, 125, 13112–13119.
(6) Gadikota, R. R.; Callam, C. S.; Wagner, T.; Del Fraino, B.; Lowary,
T. L. J. Am. Chem. Soc. 2003, 125, 4155–4165.
Episulfonium ions such as 12 have often been proposed7-15,17
to be intermediates in glycosylations of this type, and their
intermediacy provides a convenient explanation for the typically
high stereoselectivity observed. There is little evidence, however,
for their formation in these processes. Indeed, an increasing body
of work, both experimental and computational, suggests that
these glycosylations proceed through oxocarbenium ion species
such as 13, not episulfonium ions, 12. For example, all reported
computational efforts to identify stable R-oxygenated episul-
fonium ions have been unsuccessful and have suggested instead
(7) Auzanneau, F.-I.; Bundle, D. R. Carbohydr. Res. 1991, 212, 13–24.
(8) Zuurmond, H. M.; van der Klein, P. A. M.; van der Marel, G. A.; van
Boom, J. H. Tetrahedron 1993, 49, 6501–6514.
(9) Stick, R. V.; Tilbrook, D. M. G.; Williams, S. J. Aust. J. Chem. 1999,
52, 885–894.
(10) Bravo, F.; Viso, A.; Alcazar, E.; Molas, P.; Bo, C.; Castillon, S. J.
Org. Chem. 2003, 68, 686–691.
(11) Poopeiko, N.; Fernandez, R.; Barrena, M. I.; Castillon, S.; Fornies-
Camer, J.; Cardin, C. J. J. Org. Chem. 1999, 64, 1375–1379.
(12) Yang, Z.; Yu, B. Carbohydr. Res. 2001, 333, 105–114.
(13) Yu, B.; Yang, Z. Org. Lett. 2001, 3, 377–379.
(14) La´za´r, L.; Bajza, I.; Jakab, Z.; Lipta´k, A. Synlett 2005, 2242–2244.
(15) Sajtos, F.; La´za´r, L.; Borba´s, A.; Bajza, I.; Lipta´k, A. Tetrahedron
Lett. 2005, 46, 5191–5194.
(16) Nicolaou, K. C.; Mitchell, H. J.; Fylaktakidou, K. C.; Suzuki, H.;
Rodr´ıguez, R. M. Angew. Chem., Int. Ed. 2000, 39, 1089–1093.
(17) Beaver, M.; Billings, S.; Woerpel, K. Eur. J. Org. Chem. 2008, 771–
781.
9
10.1021/ja9029945 CCC: $40.75  2009 American Chemical Society
J. AM. CHEM. SOC. 2009, 131, 12937–12948 12937

A R T I C L E S
Hou et al.
Figure 1. Possible mechanistic pathways for the formation of 2-deoxy-2-thioaryl-thioglycosides from 2,3-anhdyrosugar thioarylglycosides. Pathway A
involves an episulfonium ion intermediate, whereas pathway B involves an oxocarbenium ion intermediate.
that the oxocarbenium ions are the true intermediates.10,18-20
Other studies,21,22 which report the formation of R/ꢀ mixtures
of glycosides from these reactions, also support the formation
of oxocarbenium ion intermediates, as these species would not
be expected to react in a stereospecific manner. For a more
detailed discussion of these processes, the reader is directed to
an excellent recent review by Beaver et al.17
We describe here a series of studies with thioglycoside 1,
which are aimed at understanding the mechanism by which the
migration-glycosylation shown in Figure 1 proceeds. In car-
rying out these investigations, a range of techniques was
1
employed, including low-temperature H NMR spectroscopy,
computational studies, chemical synthesis, and the measurement
Figure 2. Sulfonium ions detected by Kim et al. (ref 28) and Beaver et al.
of R-deuterium kinetic isotope effects. As discussed below in
detail, the data obtained from all approaches point to the
intermediacy of an oxocarbenium ion intermediate. Indeed, the
isotopic labeling investigations reported here provide the first
direct experimental evidence for the true nature of the reactive
intermediate in this reaction.
(ref 29) using low-temperature NMR spectroscopy.
the expected shift of the anomeric proton (H1) to lower chemical
shift compared to the parent glycosyl donors. Furthermore, an
HMBC (heteronuclear multiple bond correlation) experiment
correlated C1 and H8eq (sulfonium ion 16) and C1 and H7
(sulfonium ion 17) indicating that the SPh and SEt groups were
connected to C1.
Results and Discussion
Low-Temperature NMR Studies. In an earlier investigation5
on the mechanism of 2,3-anhydrosugar glycoside formation from
1 and 4, we used low-temperature NMR spectroscopy, as
pioneered by Crich and Sun23 and others,24-27 to probe for
reaction intermediates. Given our experience with these experi-
ments, and recent investigations by Kim et al.28 and Beaver et
al.,29 who had used NMR spectroscopy to detect the formation
of sulfonium ion intermediates in other glycosylations, we hoped
to exploit this technique to determine if a species such as 12
was formed in these reactions. In previous studies, glycosyl
donors 1428 and 1529 were activated in the absence of an alcohol,
and sulfonium ions 16 and 17, respectively, were detected by
NMR spectroscopy. These species could be characterized by
Mindful of these previous experimental studies, an NMR tube
containing a solution of compound 1 in deuterated dichlo-
romethane was cooled to -78 °C, and an H NMR spectrum
1
was recorded. The tube was then removed from the spectrom-
eter, and TMSOTf was quickly added at -78 °C. After returning
the sample to the spectrometer, the sample was warmed to room
temperature in 10° increments over 2 h, and a spectrum was
recorded at each temperature. Under these conditions no single
intermediate showing an anomeric proton at higher chemical
shift than the starting material, which based on earlier work28-30
would be indicative of an episulfonium ion, was observed.
Instead, a large number of different species were produced, the
structures of which could not be determined. Upon bringing
the spectrometer to room temperature, and allowing the tube to
stand for several hours, the product was characterized as the
hydrolyzed donor, due to absorption of atmospheric moisture.
(18) Dudley, T. J.; Smoliakova, I. P.; Hoffmann, M. R. J. Org. Chem. 1999,
64, 1247–1253.
(19) Smoliakova, I. P. Curr. Org. Chem. 2000, 4, 589–608.
(20) Jones, D. K.; Liotta, D. C. Tetrahedron Lett. 1993, 34, 7209–7212.
(21) Johnston, B. D.; Pinto, B. M. J. Org. Chem. 2000, 65, 4607–4617.
(22) Nicolaou, K.; Ladduwahetty, T.; Randall, J.; Chucholowski, A. J. Am.
Chem. Soc. 1986, 108, 2466–2467.
1
In an attempt to gain an understanding of the H NMR
spectrum that would result from an episulfonium ion derived
from 1, we explored an alternate method of generating such a
(23) Crich, D.; Sun, S. X. J. Am. Chem. Soc. 1997, 119, 11217–11223.
(24) Boebel, T. A.; Gin, D. Y. J. Org. Chem. 2005, 70, 5818–5826.
(25) Garcia, B. A.; Gin, D. Y. J. Am. Chem. Soc. 2000, 122, 4269–4279.
(26) Honda, E.; Gin, D. Y. J. Am. Chem. Soc. 2002, 124, 7343–7352.
(27) Liu, J.; Gin, D. Y. J. Am. Chem. Soc. 2002, 124, 9789–9797.
(28) Kim, J.; Yang, H.; Park, J.; Boons, G.-J. J. Am. Chem. Soc. 2005,
127, 12090–12097.
(29) Beaver, M.; Billings, S.; Woerpel, K. A. J. Am. Chem. Soc. 2008,
130, 2082–2086.
Figure 3. Alternate approach to generate an episulfonium ion.
9
12938 J. AM. CHEM. SOC. VOL. 131, NO. 36, 2009

2,3-Anhydrosugar Migration-Glycosylation Mechanism
A R T I C L E S
Figure 4. (A) Relative energy of episulfonium ions 19 and 12a, and furanosyl oxocarbenium ions 20 and 13a, as determined by B3LYP/6-311++G** DFT
calculations. (B) Relative stabilities of two possible conformers of 13a. (C) Resonance hybrid representation of 13a.
Table 1. Selected Geometrical Parameters for Structures 19, 20,
species. We reasoned that this would be possible from com-
pound 18, upon activation with tin(IV) chloride as illustrated
12a, and 13a
19
20
12a
13a
in Figure 3. Thus, 18 was synthesized (see the Supporting
Information), and then low-temperature NMR experiments
similar to those described for 1 were performed; however, the
same results were obtained. These experiments thus suggest that
a single episulfonium ion is not an intermediate in these
reactions, but further confirmation required other approaches.
r1
r2
r3
τ
1.288
1.893
2.403
1.289
1.893
2.403
1.284
1.888
2.425
1.277
1.890
2.456
90.8
90.8
92.1
93.5
Chart 2
Density Functional Theory Calculations. Density functional
theory (DFT) calculations have found increasing use in probing
glycosylation mechanisms,5,31-37 and we therefore applied this
approach to this reaction. In this case, the oxocarbenium ion
20 (Figure 4A) was minimized at the HF/6-31G** and B3LYP/
6-31+G** levels of theory. In an attempt to calculate the relative
energy difference between 20 and episulfonium ion 19, the latter
was also minimized using the same conditions. All optimizations
of 19 converged to an oxocarbenium ion as a minimum, and
no three-membered episulfonium ion species was observed. The
absence of this episulfonium ion structure can be attributed to
the stabilization of the oxocarbenium ion. In the oxocarbenium
ions observed, the lone pairs on the sulfur atom provide
stabilization of the positive charge through orbital interactions
with the π* orbital of the C1dO+ bond. Therefore, a structure
in which the positive charge is delocalized between the sulfur
atom and the oxygen in the ring may be more plausible than
one involving a strained three-membered ring.10
3
conformers (E3 and E, Figure 4B), the latter is substantially
more stable. Indeed, all attempts to optimize the E3 conformer
3
of 13a led to its conversion to the E conformation during
geometry optimization.
Our unsuccessful attempts to obtain an optimized episulfo-
nium ion structure for the determination of the relative energy
differences between 19 and 20 (as well as 12a and 13a) led us
to explore constrained geometry optimizations on these struc-
tures. Previously, Bravo et al.10 used DFT methods to determine
the geometric and electronic structure of several cations (includ-
ing episulfonium ions), which are potential intermediates in
glycosylation reactions involving 2-thioalkyl/2-thioaryl glyco-
sylating agents. In this computational study, optimized bond
lengths involving the cationic species were reported, but a
comparison of the different ring geometries (e.g., 13a-3E vs 13a-
E3) was not described. From our above calculations, bond
lengths were determined as presented in Table 1. In all cases,
we observed that the C2-S bond (r2, Figure 4) was significantly
shorter than the C1-S bond length (r3), consistent with a related
study by Bravo et al.10 As was also observed by Bravo et al.,
the C1-C2-S bond angle (τ) in these structures was ∼90°,
substantially smaller than the ideal tetrahedral geometry. A
possible rationale for this observation is that the bond angle
adjusts to place the electron-rich sulfur closer to the electron-
deficient C1dO+ group, in turn stabilizing this intermediate.
Thus, on the basis of these calculations, the intermediate 13a
cannot accurately be described as a “pure” oxocarbenium ion,
nor can it be described as a “pure” episulfonium ion. Instead,
it can be viewed as a resonance hybrid (Figure 4C), in which
the left-hand structure in the figure is the major contributor.
3
All optimizations also gave the E ring conformer as the
energetically most stable species. After B3LYP/6-311++G**
single-point energy calculations on the structures resulting from
the optimization of 19 and 20, no energy difference was
observed between the two species, and the optimized geometries
of each of these were virtually identical. The same calculations
were also done on structures 12a and 13a, and these also resulted
3
in convergence to the oxocarbenium ion in the E ring form.
Although 20 (and 13a) can adopt one of two possible envelope
(30) Dohn, D. R.; Casida, J. E. Bioorg. Chem. 1987, 15, 115–124.
(31) Zhu, X. M.; Kawatkar, S.; Rao, Y.; Boons, G.-J. J. Am. Chem. Soc.
2006, 128, 11948–11957.
(32) Berces, A.; Enright, G.; Nukada, T.; Whitfield, D. M. J. Am. Chem.
Soc. 2001, 123, 5460–5464.
(33) Ionescu, A. R.; Berces, A.; Zgierski, M. Z.; Whitfield, D. M.; Nukada,
T. J. Phys. Chem. A 2005, 109, 8096–8105.
(34) Kim, J. H.; Yang, H.; Khot, V.; Whitfield, D.; Boons, G.-J. Eur. J.
Org. Chem. 2006, 5007–5028.
(35) Yang, M. T.; Woerpel, K. A. J. Org. Chem. 2009, 74, 545–553.
(36) Whitfield, D. M.; Nukada, T. Carbohydr. Res. 2007, 342, 1291–1304.
(37) Whitfield, D. M. Carbohydr. Res. 2007, 342, 1726–1740.
9
J. AM. CHEM. SOC. VOL. 131, NO. 36, 2009 12939

A R T I C L E S
Hou et al.
Figure 5. Proposed pathway leading to 1,2-thioanhydrosugar 27 from 22 or 25.
Chart 3
In their earlier paper,10 Bravo et al. also carried out calcula-
tions on the parent episulfide compound (e.g., 12b) and showed
that the two C-S bonds had similar bond lengths (1.871 Å for
C2-S and 1.905 for C1-S). Consequently, a second set of
geometry optimizations were performed at the B3LYP/6-31G**
level of theory on 19 and 12a where the C1-S and C2-S bonds
were given fixed values corresponding to these optimal values.10
These calculations all resulted in a minimized episulfonium ion
structure, with no breakage of the C1-S bond, as was observed
in the calculations in which these bonds were not fixed. Single-
point energy calculations were then done on these optimized
structures at the B3LYP/6-311++G** level of theory. Analysis
of the newly acquired energy values for the episulfonium ions,
and comparison to the energies of the oxocarbenium ions, shows
that the oxocarbenium ions (20 and 13a) are more energetically
stable than the corresponding episulfonium ions. More specif-
ically, 20 is 6.4 kcal mol-1 more stable than 19, and 13a is 5.7
kcal mol-1 more stable than 12a. Although this approach is
admittedly artificial, it nevertheless allows us to estimate the
energy difference between the oxocarbenium and episulfonium
ion, the latter of which is unstable without the constraint. It
should be noted that initial attempts at constrained optimizations,
where the Bravo et al. optimal C-S bond lengths were not used,
resulted in an increase in the energy of the system and a halt in
the calculations. In addition, when the optimized episulfonium
ion was allowed to fully minimize, without the bond length
constraints described above, it minimized to the oxocarbeium
ion.
These computational studies are consistent with the observa-
tions from the low-temperature NMR experiments, which failed
to detect signals arising from an episulfonium ion. In contrast
to the stable [2.2.2]-bridged bicyclic sulfonium ion 17 (Figure
2), we propose that the highly strained three-member ring is
responsible for the relative instability of the episulfonium ion
compared to the oxocarbenium ion. These calculations provide
further evidence that the intermediate formed in these glyco-
sylation is a 2-thiotolyl-oxocarbenium ion, not an episulfonium
ion.
Effect of Different Aglycones. To explore further the nature
of the intermediate in this glycosylation, and to elucidate the
steric factors contributing to its stereoselectivity, we synthesized
a panel of additional donors (21-25, Chart 2) for use in the
reactions. We surmised that if the aglycone was smaller than
an aryl group (e.g., an ethyl group as in 21) there would be less
steric repulsion on the R-face, leading to an increase in the
amount of the 1,2-cis R-furanoside, should an oxocarbenium
ion be the intermediate. To probe steric effects related to the
aglycone in more detail, the t-butyl thioglycoside 22 was also
synthesized. The ability of the aglycone to stabilize a positive
charge on sulfur, as in an episulfonium ion, was to be probed
through reactions comparing the electron-withdrawing p-nitro-
phenyl thioglycoside 23 and the electron-donating p-methox-
yphenyl thioglycoside 24.
Furthermore, we envisioned that thioglycosides 22 and 25
could be used to probe the possible formation of an episulfonium
ion as illustrated in Figure 5. If an intermediate such as 26 or
28 was produced, the loss of isobutylene (from 22) or the
p-methoxylbenzyl cation (from 25) would lead to the formation
of the 1,2-thioanhydrosugar (27). Formation of this byproduct
would thus provide evidence for an episulfonium ion intermedi-
ate. An analogous approach has recently been used by Crich et
al. to probe glycosylations proposed to proceed via neighboring
group participation by acyl groups.38
Details on the synthesis of 21-25 can be found in the
Supporting Information. Once prepared, each was used in
glycosylations under the optimized conditions developed previ-
ously for benzyl alcohol (29, Chart 3) and methyl 2,3,4-tri-O-
benzoyl-R-D-mannopyranoside (30). All reactions proceeded
smoothly and provided the desired 2-thioalkyl/aryl glycosides
in good yield (Table 2).
As detailed in the table, in all but one case, the reaction was
exclusively selective for the ꢀ-isomer. In the one case in which
it was not, the glycosylation between 21 and 30 (entry 6), the
reaction still favored the ꢀ-glycoside to a large degree (R/ꢀ ratio
1
of 1:13), as determined by H NMR spectroscopy. Thus, it
appears that the steric or electronic nature of the group on sulfur
plays little, if any, role in the stereoselectivity of the reaction.
In addition, the reactions involving 22 and 25 behaved the same
(38) Crich, D.; Hu, T.; Cai, F. J. Org. Chem. 2008, 73, 8942–8953.
9
12940 J. AM. CHEM. SOC. VOL. 131, NO. 36, 2009

2,3-Anhydrosugar Migration-Glycosylation Mechanism
A R T I C L E S
Table 2. Reaction of Thioglycosides 21-25 with Alcohols 29 and
30
Scheme 1
d
ꢀ/R
entry
donor
acceptor
methoda
productb
yield (%)c
1
2
3
4
5
6
7
8
9
10
21
22
23
24
25
21
22
23
24
25
29
29
29
29
29
30
30
30
30
30
A
A
A
A
A
B
B
B
B
B
31
32
33
34
35
36
37
38
39
40
94
84
88
86
87
67
77
72
61
69
100:0
100:0
100:0
100:0
100:0
13:1
100:0
100:0
100:0
100:0
a Method A:
4 Å MS (10 equiv), CH2Cl2, reflux. Method B:
Cu(OTf)2 (1 equiv), 4 Å MS, CH2Cl2, rt. b See Chart 2 for structures of
donors and Chart 3 for structures of products. c Isolated yield after
chromatography. d Ratio determined by 1H NMR spectroscopy after
chromatography.
as the other donors, and the formation of product 27 was not
observed. It should be mentioned that the results described here
differ from earlier work by Viso et al. on a related glycosylation
in which 2-hydroxy-thioglycosides were converted to 2-deoxy-
2-thioaryl-nucleosides upon treatment under Mitsunobu condi-
tions.39 In that study, the stereoselectivity of the process was
influenced by the nature of group on the sulfur. We are unsure
as to the origin of this seemingly disparate behavior between
our work and that of Viso et al., but presumably it results from
the different reaction conditions and nucleophiles employed.
When considered with the computational and NMR investiga-
tions described above, the results detailed in Table 1 point to a
reaction governed by an SN1 pathway involving an oxocarbe-
nium ion intermediate, not one involving an episulfonium ion.
route used to prepare a suitable donor (41), which possesses
>93% deuterium at the anomeric center, is summarized in
Scheme 1.
Initially, D-arabinose (42) was oxidized to the corresponding
1,4-lactone using aqueous bromine under basic conditions. The
product was then acetylated with a catalytic amount of sulphuric
acid and acetic anhydride in acetic acid at 50 °C to yield 43 in
73% overall yield.42 The deuterium was introduced by reduction
of 43 with sodium borodeuteride.43 Subsequent acetylation of
the resulting hemiacetal with a catalytic amount of DMAP and
acetic anhydride in pyridine afforded compound 44 in 33%
overall yield. The low yield for this process was due to the
formation of the alditol, resulting from over reduction of the
lactone by sodium borodeuteride. The R-thioglycoside 45 was
obtained in 43% yield when 44 was treated with p-thiocresol
and boron trifluoride etherate in dichloromethane at 0 °C.
Deprotection of 45 was then achieved under standard conditions
to provide 46 in 85% yield. Finally, treatment with diisopro-
pylazodicarboxylate, triphenylphosphine, and benzoic acid
provided 47 in 70% yield.
For the R-DKIE measurements, a 50:50 mixture of a substrate
with and without a deuterium label at the anomeric carbon was
required. Therefore, an equal amount of 48 and 49 (obtained
by debenzoylation of 47 and 1, respectively) was combined to
give a mixture 50 containing ∼50% deuterium at C-1 (Scheme
2). In the following schemes, this material is indicated by H*
at the anomeric center. Next, 50 was acetylated with acetic acid,
or acetic acid-d3, to yield either 51 or 52 (Scheme 3). Equal
quantities of compound 51 and 52 were then mixed to give 53,
which was 50% enriched with deuterium at both C-1 and in
the acetyl protecting group.
Kinetic Isotope Effect Studies. Recently, Crich and Chan-
drasekera40 and El-Badri et al.41 reported the use of R-deuterium
kinetic isotope effects (R-DKIEs) in mechanistic investigations
of their methodology for preparing ꢀ-mannopyranosides from
either mannopyranosyl triflates or mannopyranosyl iodides,
respectively. In these studies, the glycosyl triflate or iodide was
generated and then treated with an alcohol; R-DKIEs values of
1.1-1.2 were measured. The magnitudes of these R-DKIEs
indicate that the transition state in these reactions has substantial
oxocarbenium ion character, and it was proposed that this
displacement of the leaving group (i.e., the triflate or iodide)
by the alcohol proceeded by way of an “exploded” SN2 transition
state.
The glycosylation under investigation here is more compli-
cated in that two steps are involved: the generation of an
electrophilic intermediate (e.g., either 12 or 13, Figure 1) and
its subsequent reaction with an alcohol. To probe further the
mechanism of the reaction, in particular the nature of the
transition state in the rate-determining step, we determined
R-DKIEs using a substrate labeled with deuterium at C-1. The
With 53 in hand, we measured R-DKIEs for two reactions,
the glycosylation of benzyl alcohol (29, Scheme 3) and of
methyl 2,3,4-tri-O-benzoyl-R-D-glucopyranoside (30). Each
reaction was done in triplicate. In carrying out these experiments,
donor 53 and approximately 50 mol % of internal standard
4,4,5,5-tetramethyl-2-(1-naphthyl)-1,3-dioxolane 54 were dis-
(39) Viso, A.; Poopeiko, N.; Castillo´n, S. Tetrahedron Lett. 2000, 41, 407–
411.
(40) Crich, D.; Chandrasekera, N. S. Angew. Chem., Int. Ed. 2004, 43,
5386–5389.
(41) El-Badri, M. H.; Willenbring, D.; Tantillo, D. J.; Gervay-Hague, J. J.
Org. Chem. 2007, 72, 4663–4672.
(42) Sun, R. C.; Okabe, M. Org. Syn. 1995, 72, 48–52.
(43) Bock, K.; Lundt, I.; Pedersen, C. Carbohydr. Res. 1981, 90, 7–16.
9
J. AM. CHEM. SOC. VOL. 131, NO. 36, 2009 12941

A R T I C L E S
Hou et al.
Scheme 2
Scheme 3
Table 3. R-DKIEs Measured for the Glycosylation of 29 with 53a
Table 4. R-DKIEs Measured for the Glycosylation of 30 with 53a
reaction
KIE
F
R
R0
reaction
KIE
F
R
R0
1
2
3
1.14
1.23
1.23
1.20
0.720
0.794
0.735
3.283
3.261
3.239
3.524
3.585
3.610
1
2
3
1.12
1.20
1.18
1.17
0.595
0.635
0.653
3.267
3.222
3.250
3.50
3.610
3.585
average
average
a F ) yield of 55. R ) ratio of acetate peak/anomeric peak in 55. R0
) ratio of acetate peak/anomeric peak in 53.
a F ) yield of 56. R ) ratio of acetate peak/anomeric peak in 56. R0
) ratio of acetate peak/anomeric peak in 53.
solved in chloroform-d and an 1H NMR spectrum was recorded.
The integration of acetate and the anomeric hydrogen peaks in
53 was measured against the methyl group signal of 54. After
removal of the chloroform-d, the residue was redissolved in
dichloromethane, and to this solution was added the alcohol
and the appropriate promotor (4 Å molecular sieves for 29,
copper(II) triflate for 30). After completion of the reaction, the
1H NMR spectrum of the crude reaction mixture revealed the
percentage of reaction conversion through integration of the ac-
etate methyl resonance against the methyl group signal of 54.
After flash chromatography, the 1H NMR spectrum of the
product 2-thiotolyl-2-deoxy-ꢀ-xylofuranoside (55 or 56) was
recorded and the ratio of the acetate methyl resonance against
the anomeric proton was integrated.
Therefore, when considering the two pathways outlined in
Figure 1, the formation of an electrophilic intermediate is clearly
rate-limiting. Moreover, given the magnitude of the R-DKIEs
this species must be the oxocarbenium ion 13 (pathway B), not
an episulfonium ion. These data also suggest that the formation
of 13 from 11, which involves both opening of the protonated
epoxide and 1,2-migration of the thioaryl group, is concerted
(Figure 6) and that the transition state between the species occurs
late in the reaction coordinate. It should be noted, however,
that it is impossible to determine if the intermediate is the free
oxocarbenium ion, or a structure such as 57, which is intermedi-
ate between 13 and episulfonium ion 12. In 57, the positive
charge of the oxocarbenium ion is stabilized to some degree
by the valence electrons on sulfur, a situation proposed earlier
in previous computational investigations10 and also noted in the
discussion of the DFT calculations described above.
The R-DKIE for each glycosylation was determined according
to eq 1, developed by Singleton and Thomas44
Concerted transformations of this type have not been proposed
for migration-glycosylation processes such as those under
investigation here. However, examples of similar concerted
processes do exist in the literature. For example, if one considers
an episulfonium ion and a Lewis acid activated epoxide as valid
models of each other (an admittedly crude approximation) the
transformation shown in Figure 7 suggests a concerted process
linking 11 and 13 is not unreasonable. The first, rate-limiting
step in the conversion of episulfonium ion 58 into 60 has been
proposed to be the concerted formation of cation 59.45,46 As
ln(1 - F)
ln[1 - (FR/R0)]
KIE )
(1)
where F is the fractional conversion of donor 53 (the yield of
55 or 56) and R and R0 are the ratios of the anomeric protons
integrations in the product and the starting donor, respectively,
which correspond to the D/H ratios in 55 (or 56) and 53. The
results of these calculations are shown in Tables 3 and 4; the
average R-DKIEs were determined to be 1.20 and 1.17 for 29
and 30, respectively.
These data are consistent with a rate-determining step in
which the hybridization at C-1 changes from sp3 to sp2.
(45) Lucchini, V.; Modena, G.; Pasquato, L. J. Am. Chem. Soc. 1988, 110,
6900–6901.
(46) Lucchini, V.; Modena, G.; Pasquato, L. J. Am. Chem. Soc. 1991, 113,
(44) Singleton, D.; Thomas, A. J. Am. Chem. Soc. 1995, 117, 9357–9358.
6600–6607.
9
12942 J. AM. CHEM. SOC. VOL. 131, NO. 36, 2009

2,3-Anhydrosugar Migration-Glycosylation Mechanism
A R T I C L E S
Figure 6. Proposed pathway leading from 11 to oxocarbenium ion 13 or a related intermediates 57; although all the data reported in this paper suggests that
episulfonium ion 12 is not produced in these reactions, it is shown for purposes of comparison.
Figure 7. Proposed mechanism for the conversion of 57 to 59, via concerted formation of 58 (refs 45 and 46).
through a late transition state, comparable to a structure of the
type 57. Thus, the magnitude of the R-DKIE measured for
reactions of 1 is consistent with the process outlined in Figure
6. It should be appreciated that our measurements do not rule
out the fast formation of an episulfonium ion followed by a
rate-determining rearrangement to the oxocarbenium ion, as has
been suggested previously.20 However, with regard to the
stereocontrol of the glycosylation reaction under investigation
here, this mechanistic difference is immaterial.
It should also be mentioned that this isotope effect clearly
rules out attack of the oxygen on the electrophile as the rate-
determining step. In the case of rate-limiting SN1 attack of the
alcohol on 13 one would expect an inverse R-DKIE because
the C-1 hybridization would change from sp2 to sp3. Similarly,
were the SN2 attack of the alcohol on an episulfonium ion such
as 12 rate-limiting an inverse R-DKIEs would also be expected,
as has been calculated for the methanolysis of protonated
epoxides proceeding via tight SN2-like transition states.51
Figure 8. De Groot rearrangement (refs 48 and 49).
part of the mechanistic investigations on the formation of 60
from 58, substrates in which the methyl groups were deuterated
(e.g., 61) were synthesized, and a kH/kD of 1.27 was measured.46
The magnitude of this R-DKIE is in agreement with a concerted
process possessing a transition state with substantial positive
charge.
The similarity between this R-DKIE and those measured for
the reactions of 1 support a concerted process in the formation
of 13 from 11. Furthermore, inspection of 2,3-anhydrosugar
thioglycosides for which X-ray crystallographic data is avail-
able47 indicates that in these rigid species the C-S bond is
aligned in a near-perfect antiperiplanar orientation with both
the C2-Oep bond, as well as the region of space expected to be
occupied by one of the ring oxygen lone pairs. Such an
arrangement would facilitate a concerted 1,2-shift concomitant
with epoxide ring-opening leading to 13.
Notwithstanding the example described in the previous
paragraphs, there is a paucity of R-DKIE data on analogous
reactions with which the data in Tables 3 and 4 can be
compared. For example, none of the migration-glycosylation
processes reported previously have been subjected to such
analysis, nor have similar transformations outside the glycosy-
lation area, such as the de Groot rearrangement (Figure 8)48,49
and related variants.50 However, R-DKIEs obtained for the acid-
catalyzed opening of epoxides can prove instructive. In a study
published in 1980,51 Hanzlik and Westkaemper measured the
effect of isotopic substitution on the methanolysis of p-
nitrostryene epoxide under acidic conditions. These studies
demonstrated that deuterium labeling of the “reacting” carbon
(analogous to C-1 in 12, Figure 6) showed an estimated R-DKIE
of 1.19 and on that basis proposed that the reaction proceeded
Explanation for Glycosylation Selectivity. All of the data
presented above suggests that the glycosylation-migration
reaction of 1 proceeds via concerted rate-determining formation
of oxocarbenium ion intermediate 13. No evidence, from either
experimental or computational studies, could be obtained for
the formation of an episulfonium ion intermediate. It should be
noted, however, that the computational studies do suggest that
some stabilization of the positive charge by the sulfur atom is
possible via a structure similar to 57 (Figure 6). A model
rationalizing the observed highly favored ꢀ-face approach of
the acceptor is presented in Figure 9. The furanosyl oxocarbe-
nium ion, in which the bond linking C-1 and O-5 has significant
double-bond character can adopt two possible conformations,
62 and 63, the E3 and 3E conformers, respectively. The
computational studies described above have indicated that the
3
most stable conformation of this oxocarbenium ion is the E
conformer 63. This is presumably due to not only favorable
stereoelectronic and electrostatic effects35,52 between the sub-
stituents at C-2 and C-3 and the positively charged ring oxygen
in 63 but also due to unfavorable steric interactions between
the benzoyloxymethyl group at C-4 and H-2 in 62. We propose
that based on Woerpel’s “inside attack” model,53,54 that nu-
cleophilic attack from the top face of the ring should be favored,
giving the 1,2-trans glycoside as the major product. Formation
of the 1,2-cis product would require attack of the alcohol on
(47) Gallucci, J. C.; Gadikota, R. R.; Lowary, T. L. Acta Crystallogr., Sect.
C 2000, 56, E365–E365.
(48) Jansen, B. J. M.; Peperzak, R. M.; Degroot, A. Recl. TraV. Chim. Pays-
Bas 1987, 106, 549–553.
(49) Jansen, B. J. M.; Peperzak, R. M.; Degroot, A. Recl. TraV. Chim. Pays-
Bas 1987, 106, 489–494.
(50) Sromek, A.; Gevorgyan, V. Top. Curr. Chem. 2007, 274, 77–124.
(51) Hanzlik, R.; Westkaemper, R. J. Am. Chem. Soc. 1980, 102, 2464–
2467.
(52) Wolfe, S. Acc. Chem. Res. 1972, 5, 102–111.
9
J. AM. CHEM. SOC. VOL. 131, NO. 36, 2009 12943

A R T I C L E S
Hou et al.
Figure 9. Proposed model for stereoselectivity in migration/glycosylation reactions with 1.
(s, 1H, H-1), 4.84 (d, 1H, J ) 11.6 Hz, PhCH2), 4.72-4.67 (m,
2H, H-4, H-5a), 4.59-4.53 (m, 2H, H-5b, PhCH2), 4.27 (dd, 1H,
J3,OH ) 10.6 Hz, J3,4 ) 3.5 Hz, H-3), 3.40 (s, 1H, H-2), 3.20 (d,
1H, J3,OH ) 10.6 Hz, OH), 2.65 (q, 2H, J ) 7.3 Hz, CH2CH3),
1.30 (t, 3H, J ) 7.3 Hz, CH2CH3); 13C NMR (125 MHz, CDCl3,
δC) 166.4 (CdO), 136.8 (Ar), 133.1 (2 × Ar), 130.0 (Ar), 129.8
(2 × Ar), 128.6 (2 × Ar), 128.4 (2 × Ar), 128.2 (2 × Ar), 107.0
(C-1), 81.1 (C-4), 76.6 (C-3), 69.8 (PhCH2), 64.5 (C-5), 55.2 (C-
2), 26.2 (CH2CH3), 14.7 (CH2CH3). HRMS (ESI) calcd for (M +
Na) C21H24O5S: 411.1237. Found: 411.1242.
Benzyl 5-O-Benzoyl-2-t-butylthio-2-deoxy-ꢀ-D-xylofuranoside
(32). Benzyl alcohol (29, 35 mg, 0.33 mmol) and 22 (67 mg, 0.22
mmol) were coupled in CH2Cl2 (15 mL) using 4 Å molecular sieves
(1.0 g) as described for the preparation of 31. Chromatography (5:1
hexane-EtOAc) of the crude product gave 32 (75 mg, 84%) as a
colorless oil: Rf 0.81 (2:1 hexane-EtOAc); [R]D -18.9 (c, 0.1,
CH2Cl2); 1H NMR (500 MHz, CDCl3, δH) 8.11-8.07 (m, 2H, Ar),
7.59-7.53 (m, 1H, Ar), 7.46-7.41 (m, 2H, Ar), 7.38-7.28 (m,
5H), 5.19 (s, 1H, H-1), 4.82 (d, 1H, J ) 11.6 Hz, PhCH2),
4.71-4.63 (dd, 1H, J5a,5b ) 11.9 Hz, J4,5a ) 7.7 Hz, H-5a),
4.60-4.51 (m, 3H, PhCH2, H-4, H-5b), 4.28 (dd, 1H, J3,OH ) 11.2
the bottom face of the less populated conformer 62 or attack
cis to the bulky thiotolyl group in the dominant conformer 63.
In conclusion, a series of experiments probing the mechanism
of the 2,3-anhydrosugar migration-glycosylation reaction were
performed on thioglycoside 1. All of the data support the
pathway shown in Figure 9, which involves the rate-determining
generation of an oxocarbenium ion intermediate that reacts in
a stereoselective manner with an alcohol nucleophile. Particu-
larly compelling evidence comes from the measured R-DKIEs
(magnitudes of 1.17-1.20), which provide, for the first time, a
direct measure of a rate-limiting transition state in migration-
glycosylations of this type. Thus, although the formation of an
episulfonium intermediate provides a convenient explanation
for the observed stereospecificity of these reactions, we propose
instead that the stereochemical outcome of the reaction results
from “inside attack”53,54 of the nucleophile onto the lowest
energy conformer of the oxocarbenium ion.
Although these investigations have been performed only for
thioglycoside 1, in light of previous studies described above,
we consider it likely that similar reactions of thiofuranoside 4
and pyranosides 6 and 7 also proceed via a similar pathway.
More generally, the results reported here, when considered
together with those reported previously by other groups,10,18-22
may suggest that all glycosylations7-17 involving potential
equilibration of glycosyl 1,2-episulfonium and 2-thioalkyl/aryl
oxocarbenium ions, involve the latter intermediates. Further
support for this assertion is of course necessary, and the
measurement of R-DKIEs for similar processes will no doubt
provide essential insight into the mechanisms of these reactions.
Hz, J3,4 ) 2.7 Hz, H-3), 3.33 (s, 1H, H-2), 3.25 (d, 1H, J3,OH
)
11.2 Hz, OH), 1.39 (s, 9H, C(CH3)3); 13C NMR (125 MHz, CDCl3,
δC) 166.4 (CdO), 136.9 (Ar), 133.01 (Ar), 133.03 (Ar), 129.8 (Ar
× 2), 128.6 (Ar × 2), 128.3 (Ar × 2), 128.2 (Ar × 2), 128.1 (Ar),
108.5 (C-1), 81.1 (C-4), 78.3 (C-3), 69.8 (PhCH2), 64.5 (C-5), 52.9
(C-2), 44.2 (SC(CH3)3), 31.2 (SC(CH3)3). HRMS (ESI) calcd for
(M + Na) C23H28O5S: 439.1550. Found: 439.1548.
Benzyl 5-O-Benzoyl-2-deoxy-2-p-methoxyphenylthio-ꢀ-D-xy-
lofuranoside (33). Benzyl alcohol (29, 48 mg, 0.45 mmol) and 23
(80 mg, 0.22 mmol) were coupled in CH2Cl2 (15 mL) with 4 Å
molecular sieves (1.5 g) as described for the preparation of 31.
The mixture was heated at reflux until TLC showed completion of
the reaction. Chromatography (5:1 hexane-EtOAc) of the crude
product gave 33 (92 mg, 88%) as a colorless oil: Rf 0.65 (2:1
hexane-EtOAc); [R]D -21.6 (c, 0.4, CH2Cl2); 1H NMR (500 MHz,
CDCl3, δH) 8.11-8.07 (m, 2H, Ar), 7.60-7.54 (m, 1H, Ar),
7.47-7.25 (m, 9H, Ar), 6.88-6.84 (m, 2H, Ar), 5.16 (s, 1H, H-1),
4.81 (d, 1H, J ) 11.6 Hz, PhCH2), 4.78-4.73 (m, 1H, H-4), 4.69
(dd, 1H, J5a,5b ) 11.9 Hz, J4,5a ) 4.5 Hz, H-5a), 4.56 (dd, 1H, J5a,5b
) 11.9 Hz, J4,5b ) 7.4 Hz, H-5b), 4.52 (d, 1H, J ) 11.6 Hz, PhCH2),
4.25 (dd, 1H, J3,OH ) 11.2 Hz, J3,4 ) 4.3 Hz, H-3), 3.81 (s, 3H,
Experimental Section
Benzyl 5-O-Benzoyl-2-deoxy-2-ethylthio-ꢀ-D-xylofuranoside
(31). To a solution of benzyl alcohol (29, 51 mg, 0.47 mmol) and
21 (88 mg, 0.31 mmol) in CH2Cl2 (15 mL) was added 4 Å
molecular sieves (1.50 g). The mixture was heated at reflux until
TLC showed completion of the reaction. The reaction mixture was
concentrated to a crude residue that was purified by chromatography
(5:1 hexane-EtOAc) to afford 31 (114 mg, 94%) as a colorless
oil: Rf 0.62 (2:1 hexane-EtOAc); [R]D -24.1 (c, 1.0, CH2Cl2); 1H
NMR (500 MHz, CDCl3, δH) 8.11-8.07 (m, 2H, Ar), 7.59-7.53
(m, 1H, Ar), 7.46-7.41 (m, 2H, Ar), 7.38-7.29 (m, 5H, Ar), 5.16
OCH3), 3.70 (s, 1H, H-2), 3.14 (d, 1H, J3,OH ) 11.2 Hz, OH); 13
C
NMR (125 MHz, CDCl3, δC) 166.4 (CdO), 160.0 (Ar), 136.8 (Ar),
134.8 (2 × Ar), 133.1 (Ar), 130.0 (Ar), 129.8 (2 × Ar), 128.6 (2
× Ar), 128.4 (2 × Ar), 128.1 (2 × Ar), 128.0 (Ar), 123.1 (Ar),
115.0 (2 × Ar), 106.2 (C-1), 81.2 (C-4), 75.8 (C-3), 69.8 (PhCH2),
64.5 (C-5), 59.1 (OCH3), 55.4 (C-2). HRMS (ESI) calcd for (M +
Na) C26H26O6S: 489.1342. Found: 489.1345.
(53) Smith, D. M.; Tran, M. B.; Woerpel, K. A. J. Am. Chem. Soc. 2003,
125, 14149–14152.
(54) Larsen, C. H.; Ridgway, B. H.; Shaw, J. T.; Woerpel, K. A. J. Am.
Chem. Soc. 1999, 121, 12208–12209.
9
12944 J. AM. CHEM. SOC. VOL. 131, NO. 36, 2009

2,3-Anhydrosugar Migration-Glycosylation Mechanism
A R T I C L E S
Benzyl 5-O-Benzoyl-2-deoxy-2-p-nitrophenylthio-ꢀ-D-xylo-
furanoside (34). Benzyl alcohol (29, 28 mg, 0.26 mmol) and 24
(64 mg, 0.17 mmol) were coupled in CH2Cl2 (15 mL) with 4 Å
molecular sieves (1.0 g) as described for the preparation of 31.
Chromatography (5:1 hexane-EtOAc) of the crude product gave
34 (71 mg, 86%) as a colorless oil: Rf 0.68 (2:1 hexane-EtOAc);
[R]D -3.0 (c, 0.3, CH2Cl2); 1H NMR (500 MHz, CDCl3, δH)
8.15-8.05 (m, 4H, Ar), 7.60-7.55 (m, 1H, Ar), 7.46-7.42 (m,
2H, Ar), 7.39-7.30 (m, 7H, Ar), 5.19 (s, 1H, H-1), 4.86 (d, 1H, J
) 11.6 Hz, PhCH2), 4.75-4.70 (m, 2H, H-4, H-5a), 4.61-4.56
(m, 2H, H-5b, PhCH2), 4.34 (dd, 1H, J3,OH ) 10.9 Hz, J3,4 ) 5.1
Hz, H-3), 3.99 (s, 1H, H-2), 3.30 (d, 1H, J3,OH ) 10.9 Hz, OH);
13C NMR (125 MHz, CDCl3, δC) 166.4 (CdO), 145.8 (Ar), 144.1
(Ar), 136.3 (Ar), 133.2 (2 × Ar), 129.78 (Ar), 129.75 (2 × Ar),
128.7 (2 × Ar), 128.43 (2 × Ar), 128.36 (2 × Ar), 127.3 (2 ×
Ar), 124.3 (2 × Ar), 105.6 (C-1), 81.2 (C-4), 75.8 (C-3), 70.2
(PhCH2), 63.9 (C-5), 56.0 (C-2). HRMS (ESI) calcd for (M + Na)+
C25H23NO7S: 504.1088. Found: 504.1081.
Methyl 5-O-Benzoyl-2-t-butylthio-2-deoxy-ꢀ-D-xylofuranosyl-
(1f6)-2,3,4-tri-O-benzoyl-r-D-mannopyranoside (37). Alcohol
30 (130 mg, 0.26 mmol) and 22 (79 mg, 0.26 mmol) were coupled
in CH2Cl2 (15 mL) containing 4 Å molecular sieves (200 mg) with
Cu(OTf)2 (93 mg, 0.26 mmol) as described for the preparation of
36. Chromatography (4:1 hexane-EtOAc) of the crude product gave
37 (162 mg, 77%) as a colorless oil: Rf 0.52 (2:1 hexane-EtOAc);
1
[R]D -62.6 (c, 0.5, CH2Cl2); H NMR (600 MHz, CDCl3, δH)
8.13-8.11 (m, 2H, Ar), 8.05-8.02 (m, 2H, Ar), 7.96-7.93 (m,
2H, Ar), 7.83-7.80 (m, 2H, Ar), 7.62-7.58 (m, 1H, Ar), 7.56-7.47
(m, 4H, Ar), 7.44-7.34 (m, 5H, Ar), 7.27-7.24 (m, 2H, Ar), 5.96
(dd, 1H, J4,5 ) J3,4 ) 10.0 Hz, H-4), 5.85 (dd, 1H, J3,4 ) 10.0 Hz,
J2,3 ) 3.3 Hz, H-3), 5.71 (dd, 1H, J2,3 ) 3.3 Hz, J1,2 ) 1.8 Hz,
H-2), 5.15 (s, 1H, H-1′), 4.99 (s, 1H, H-1), 4.67-4.55 (m, 3H,
H-4′, H-5a′, H-5b′), 4.28 (dd, 1H, J3′,OH ) 11.9 Hz, J3′,4′ ) 4.1 Hz,
H-3′), 4.23 (ddd, 1H, J4,5 ) 10.0 Hz, J5,6a ) J5,6b ) 3.4 Hz, H-5),
4.11 (dd, 1H, J6a,6b ) 10.5 Hz, J5,6a ) 3.4 Hz, H-6a), 3.64 (d, 1H,
J3′,OH ) 11.9 Hz, OH), 3.62 (dd, 1H, J6a,6b ) 10.5 Hz, J5,6b ) 3.4
Hz, H-6b), 3.55 (s, 3H, OCH3), 3.39 (s, 1H, H-2′); 13C NMR (125
MHz, CDCl3, δC) 166.3 (CdO), 165.4 (2 × CdO), 165.4 (CdO),
133.44 (Ar), 133.39 (Ar), 133.1 (Ar), 132.9 (Ar), 130.07 (Ar),
130.05 (2 × Ar), 129.74 (2 × Ar), 129.72 (2 × Ar), 129.3 (Ar),
129.2 (Ar), 129.1 (Ar), 128.6 (2 × Ar), 128.5 (2 × Ar), 128.41 (2
× Ar), 128.35 (2 × Ar), 128.2 (2 × Ar), 108.7 (C-1′), 98.9 (C-1),
81.5 (C-4′), 78.4 (C-3′), 70.2 (C-3), 70.1 (C-2), 69.4 (C-5), 66.7
(C-4), 64.7 (C-5′), 64.5 (C-6), 55.7 (OCH3), 53.0 (C-2′), 44.2
(SC(CH3)3), 31.1 (SC(CH3)3). HRMS (ESI) calcd for (M + Na)
C44H46O13S: 837.2551. Found: 837.2544.
Methyl 5-O-Benzoyl-2-deoxy-2-p-methoxyphenylthio-ꢀ-D-xy-
lofuranosyl-(1f6)-2,3,4-tri-O-benzoyl-r-D-mannopyranoside (38).
Alcohol 30 (148 mg, 0.29 mmol) and 23 (105 mg, 0.29 mmol)
were coupled in CH2Cl2 (15 mL) containing 4 Å molecular sieves
(250 mg) using Cu(OTf)2 (106 mg, 0.29 mmol) as described for
the preparation of 36. Chromatography (4:1 hexane-EtOAc) of
the crude product gave 38 (182 mg, 72%) as a colorless oil: Rf
0.43 (2:1 hexane-EtOAc); [R]D -56.0 (c, 0.49, CH2Cl2); 1H NMR
(500 MHz, CDCl3, δH) 8.10-8.03 (m, 4H, Ar), 7.96-7.92 (m, 2H,
Ar), 7.83-7.80 (m, 2H, Ar), 7.60-7.47 (m, 3H, Ar), 7.46-7.34
(m, 8H, Ar), 7.28-7.24 (m, 3H, Ar), 6.90-6.86 (m, 2H, Ar), 5.93
(dd, 1H, J4,5 ) J3,4 ) 10.0 Hz, H-4), 5.85 (dd, 1H, J3,4 ) 10.0 Hz,
J2,3 ) 3.3 Hz, H-3), 5.70 (dd, 1H, J2,3 ) 3.3 Hz, J1,2 ) 1.8 Hz,
H-2), 5.15 (s, 1H, H-1′), 4.96 (d, 1H, J1,2 ) 1.8 Hz, H-1), 4.80-4.75
(m, 1H, H-4′), 4.68-4.60 (m, 2H, H-5a′, H-5b′), 4.27-4.18 (m,
2H, H-3′, H-5), 4.11 (dd, 1H, J6a,6b ) 10.9 Hz, J5,6a ) 2.4 Hz,
H-6a), 3.82 (s, 3H, ArOCH3), 3.78 (s, 1H, H-2′), 3.61 (dd, 1H,
J6a,6b ) 10.9 Hz, J5,6b ) 3.9 Hz, H-6b), 3.52 (s, 3H, OCH3); 13C
NMR (125 MHz, CDCl3, δC) 166.3 (CdO), 165.4 (3 × CdO),
159.9 (Ar), 134.7 (2 × Ar), 133.41 (Ar), 133.40 (Ar), 133.1 (Ar),
133.0 (Ar), 130.1 (Ar), 130.0 (2 × Ar), 129.8 (2 × Ar), 129.7 (2
× Ar), 129.3 (Ar), 129.2 (Ar), 129.0 (Ar), 128.52 (2 × Ar), 128.45
(2 × Ar), 128.31 (2 × Ar), 128.28 (2 × Ar), 128.2 (2 × Ar), 123.3
(Ar), 115.0 (2 × Ar), 106.4 (C-1′), 98.8 (C-1), 81.6 (C-4′), 75.9
(C-3′), 70.12 (C-3), 70.08 (C-2), 69.3 (C-5), 66.8 (C-4), 64.7 (C-
5′), 64.4 (C-6), 59.1 (C-2′), 55.7 (ArOCH3), 55.4 (OCH3). HRMS
(ESI) calcd for (M + Na) C47H44O14S: 887.2344. Found: 887.2350.
Methyl 5-O-Benzoyl-2-deoxy-2-p-nitrophenylthio-ꢀ-D-xylo-
furanosyl-(1f6)-2,3,4-tri-O-benzoyl-r-D-mannopyranoside (39).
Alcohol 30 (143 mg, 0.28 mmol) and 24 (105 mg, 0.28 mmol)
were coupled in CH2Cl2 (15 mL) containing 4 Å molecular sieves
(250 mg) with Cu(OTf)2 (102 mg, 0.28 mmol) as described for the
synthesis of 36. Chromatography (4:1 hexane-EtOAc) of the crude
product gave 39 (152 mg, 61%) as a colorless oil: Rf 0.49 (2:1,
hexane-EtOAc); [R]D -56.4 (c, 0.8, CH2Cl2); 1H NMR (500 MHz,
CDCl3, δH) 8.20-8.16 (m, 2H, Ar), 8.14-8.11 (m, 2H, Ar),
8.06-8.02 (m, 2H, Ar), 7.99-7.96 (m, 2H, Ar), 7.84-7.81 (m,
2H, Ar), 7.66-7.61 (m, 1H, Ar), 7.58-7.37 (m, 11H, Ar),
7.30-7.25 (m, 2H, Ar), 6.05 (dd, 1H, J4,5 ) J3,4 ) 10.1 Hz, H-4),
5.89 (dd, 1H, J3,4 ) 10.1 Hz, J2,3 ) 3.3 Hz, H-3), 5.72 (dd, 1H, J2,3
) 3.3 Hz, J1,2 ) 1.8 Hz, H-2), 5.20 (s, 1H, H-1′), 5.03 (d, 1H, J1,2
Benzyl 5-O-Benzoyl-2-deoxy-2-p-methoxybenzylthio-ꢀ-D-xy-
lofuranoside (35). Benzyl alcohol (29, 22 mg, 0.20 mmol) and 25
(50 mg, 0.13 mmol) were coupled in CH2Cl2 (15 mL) with 4 Å
molecular sieves (0.75 g). Chromatography (4:1 hexane-EtOAc)
of the crude product gave 35 (56 mg, 87%) as a colorless oil: Rf
0.67 (2:1 hexane-EtOAc); [R]D -21.9 (c, 0.1, CH2Cl2); 1H NMR
(500 MHz, CDCl3, δH) 8.09-8.05 (m, 2H, Ar), 7.58-7.55 (m, 1H,
Ar), 7.45-7.41 (m, 2H, Ar), 7.37-7.28 (m, 5H, Ar), 7.22-7.19
(m, 2H, Ar), 6.87-6.83 (m, 2H, Ar), 5.05 (s, 1H, H-1), 4.77 (d,
1H, J ) 11.6 Hz, PhCH2), 4.68-4.63 (m, 2H, H-4, H-5a), 4.53
(dd, 1H, J5a,5b ) 11.7 Hz, J4,5b ) 8.3 Hz, H-5b), 4.46 (d, 1H, J )
11.6 Hz, PhCH2), 4.19 (d, 1H, J3,OH ) 11.1 Hz, J3,4 ) 4.3 Hz,
H-3), 3.80 (s, 3H, OCH3), 3.77 (s, 2H, ArCH2S), 3.26 (s, 1H, H-2),
3.07 (d, 1H, J3,OH ) 11.1 Hz, OH); 13C NMR (125 MHz, CDCl3,
δC) 166.4 (CdO), 159.0 (Ar), 136.8 (Ar), 133.0 (Ar), 130.02 (2 ×
Ar), 129.99 (Ar), 129.7 (2 × Ar), 129.1 (Ar), 128.6 (2 × Ar), 128.4
(2 × Ar), 128.2 (2 × Ar), 128.1 (Ar), 114.1 (2 × Ar), 106.7 (C-1),
81.2 (C-4), 76.3 (C-3), 69.8 (PhCH2), 64.4 (C-5), 55.3 (OCH3),
54.9 (C-2), 36.0 (CH2S). HRMS (ESI) calcd for (M + Na)
C27H28O6S: 503.1499. Found: 503.1496.
Methyl 5-O-Benzoyl-2-deoxy-2-ethylthio-ꢀ-D-xylofuranosyl-
(1f6)-2,3,4-tri-O-benzoyl-r-D-mannopyranoside (36). To a solu-
tion of 30 (128 mg, 0.25 mmol) and 21 (71 mg, 0.25 mmol) in
CH2Cl2 (15 mL) was added 4 Å molecular sieves (200 mg). After
the mixture was stirred at room temperature for 30 min, Cu(OTf)2
(92 mg, 0.25 mmol) was added and the solution was stirred for an
additional 2 h. After neutralization with triethylamine, the reaction
mixture was concentrated to a crude residue that was purified by
chromatography (4:1 hexane-EtOAc) to afford 36 (133 mg, 67%)
as a colorless oil: Rf 0.56 (2:1 hexane-EtOAc); [R]D -77.9 (c,
0.7, CH2Cl2); 1H NMR (500 MHz, CDCl3, δH) 8.13-8.09 (m, 2H,
Ar), 8.05-8.02 (m, 2H, Ar), 7.96-7.94 (m, 2H, Ar), 7.83-7.80
(m, 2H, Ar), 7.63-7.35 (m, 10H, Ar), 7.28-7.24 (m, 2H, Ar), 5.94
(dd, 1H, J4,5 ) J3,4 ) 10.1 Hz, H-4), 5.85 (dd, 1H, J3,4 ) 10.1 Hz,
J2,3 ) 3.3 Hz, H-3), 5.70 (dd, 1H, J2,3 ) 3.3 Hz, J1,2 ) 1.6 Hz,
H-2), 5.13 (s, 1H, H-1′), 4.99 (d, 1H, J1,2 ) 1.6 Hz, H-1), 4.70-4.59
(m, 3H, H-4′, H-5a′, H-5b′), 4.28-4.20 (m, 2H, H-5, H-3′), 4.13
(dd, 1H, J6a,6b ) 10.9 Hz, J5,6a ) 2.2 Hz, H-6a), 3.62 (dd, 1H, J6a,6b
) 10.9 Hz, J5,6b ) 3.6 Hz, H-6b), 3.57 (d, 1H, J3′,OH ) 12.0 Hz,
OH), 3.55 (s, 3H, OCH3), 3.46 (s, 1H, H-2′), 2.69 (q, 2H, J ) 7.3
Hz, CH2CH3), 1.33 (t, 3H, J ) 7.3 Hz, CH2CH3); 13C NMR (125
MHz, CDCl3, δC) 166.3 (CdO), 165.44 (2 × CdO), 165.42 (CdO),
133.44 (Ar), 133.43 (Ar), 133.1 (Ar), 133.0 (Ar), 130.0 (2 × Ar),
129.8 (2 × Ar), 129.73 (2 × Ar), 129.72 (2 × Ar), 129.3 (Ar),
129.2 (Ar), 129.0 (2 × Ar), 128.52 (2 × Ar), 128.46 (2 × Ar),
128.3 (2 × Ar), 128.2 (2 × Ar), 107.3 (C-1′), 98.9 (C-1), 81.5
(C-4′), 76.6 (C-3′), 70.2 (C-3), 70.1 (C-2), 69.3 (C-5), 66.8 (C-4),
64.6 (C-5′), 64.5 (C-6), 55.7 (OCH3), 55.4 (C-2′), 26.3 (CH2CH3),
14.7 (CH2CH3). HRMS (ESI) calcd for (M + Na) C42H42O13S:
809.2238. Found: 809.2225.
9
J. AM. CHEM. SOC. VOL. 131, NO. 36, 2009 12945

A R T I C L E S
Hou et al.
) 1.8 Hz, H-1), 4.76-4.66 (m, 3H, H-4′, H-5a′, H-5b′), 4.34 (dd,
(CH3CO), 20.3 (CH3CO). HRMS (ESI) calcd for (M + Na)
C11H14O8: 297.0581. Found: 297.0583.
1H, J3′,OH ) 11.9 Hz, J3′,4′ ) 4.4 Hz, H-3′), 4.24 (ddd, 1H, J4,5
)
10.1 Hz, J5,6a ) J5,6b ) 2.7 Hz, H-5), 4.20 (dd, 1H, J6a,6b ) 10.7
Hz, J5,6a ) 2.7 Hz, H-6a), 4.09 (s, 1H, H-2′), 3.78 (d, 1H, J3′,OH
1,2,3,5-Tetra-O-acetyl-D-[1-D]-arabinofuranose (44). A solu-
tion of 43 (3.9 g, 14.22 mmol) in CH3OH (50 mL) was cooled to
0 °C and stirred with Amberlite IR-120 (H+) resin (10 g). Sodium
borodeuteride (NaBD4, 0.60 g, 14.22 mmol) was added in portion.
The reaction mixture was stirred for an additional 30 min at 0 °C
before the resin was filtered and then washed with CH3OH. The
filtrate was concentrated to a crude residue that was purified by
chromatography (2:1 hexane-EtOAc) to afford 2,3,5-tri-O-acetyl-
D-[1-2H]-arabinofuranose (1.5 g, 38%) as a colorless oil. This oil
(1.5 g, 5.41 mmol) was redissolved in pyridine (10 mL). To this
solution was added Ac2O (1.02 mL, 10.82 mmol) and DMAP (75
mg, 0.27 mmol). The reaction mixture was stirred at room
temperature for 16 h, the pyridine was evaporated at 50 °C, and
thecrudeproductwaspurifiedbychromatography(2:1hexane-EtOAc)
)
11.9 Hz, OH), 3.62 (dd, 1H, J6a,6b ) 10.7 Hz, J5,6b ) 2.7 Hz, H-6b),
3.56 (s, 3H, OCH3); 13C NMR (125 MHz, CDCl3, δC) 166.3 (CdO),
165.6 (CdO), 165.43 (CdO), 165.40 (CdO), 145.8 (Ar), 144.2
(Ar), 133.6 (Ar), 133.5 (Ar), 133.2 (Ar), 133.1 (Ar), 130.0 (2 ×
Ar), 129.9 (Ar), 129.8 (2 × Ar), 129.7 (4 × Ar), 129.4 (Ar), 129.1
(Ar), 128.9 (Ar), 128.57 (2 × Ar), 128.55 (2 × Ar), 128.4 (2 ×
Ar), 128.3 (2 × Ar), 127.1 (2 × Ar), 124.4 (2 × Ar), 105.8 (C-1′),
99.0 (C-1), 81.8 (C-4′), 75.8 (C-3′), 70.2 (C-2), 70.0 (C-3), 69.0
(C-5), 66.6 (C-4), 64.22 (C-5′), 64.17 (C-6), 56.0 (C-2′), 55.8
(OCH3). HRMS (ESI) calcd for (M + Na)+ C46H41NO15S: 902.2089.
Found: 902.2094.
Methyl 5-O-Benzoyl-2-deoxy-2-p-methoxybenzylthio-ꢀ-D-xy-
lofuranosyl-(1f6)-2,3,4-tri-O-benzoyl-r-D-mannopyranoside (40).
Alcohol 30 (129 mg, 0.25 mmol) and 25 (95 mg, 0.25 mmol) were
coupled in CH2Cl2 (15 mL) containing 4 Å molecular sieves (250
mg) with Cu(OTf)2 (92 mg, 0.25 mmol) as described for the
synthesis of 36. Chromatography (4:1 hexane-EtOAc) of the crude
product gave 40 (155 mg, 69%) as a colorless oil: Rf 0.47 (2:1
hexane-EtOAc); [R]D -67.5 (c, 0.8, CH2Cl2); 1H NMR (500 MHz,
CDCl3, δH) 8.08-8.01 (m, 4H, Ar), 7.99-7.95 (m, 2H, Ar),
7.84-7.80 (m, 2H, Ar), 7.61-7.50 (m, 3H, Ar), 7.45-7.35 (m,
7H, Ar), 7.30-7.24 (m, 4H, Ar), 6.84-6.80 (m, 2H, Ar), 5.96 (dd,
1H, J4,5 ) J3,4 ) 10.0 Hz, H-4), 5.85 (dd, 1H, J3,4 ) 10.0 Hz, J2,3
) 3.3 Hz, H-3), 5.71 (dd, 1H, J2,3 ) 3.3 Hz, J1,2 ) 1.8 Hz, H-2),
5.06 (s, 1H, H-1′), 4.96 (d, 1H, J1,2 ) 1.8 Hz, H-1), 4.66-4.57 (m,
3H, H-4′, H-5a′, H-5b′), 4.22-4.14 (m, 2H, H-5, H-3′), 4.09 (dd,
1H, J6a,6b ) 10.9 Hz, J5,6a ) 2.6 Hz, H-6a), 3.86 (d, 1H, J ) 13.4
Hz, ArCH2S), 3.82 (d, 1H, J ) 13.4 Hz, ArCH2S), 3.70 (s, 3H,
ArOCH3), 3.57 (dd, 1H, J6a,6b ) 10.9 Hz, J5,6b ) 3.6 Hz, H-6b),
3.52 (s, 3H, OCH3), 3.50 (d, 1H, J3′,OH ) 8.1 Hz, OH), 3.40 (s,
1H, H-2′); 13C NMR (125 MHz, CDCl3, δC) 166.3 (CdO), 165.45
(CdO), 165.41 (CdO), 165.40 (CdO), 158.9 (Ar), 133.44 (Ar),
133.41 (Ar), 133.1 (Ar), 133.0 (Ar), 130.1 (2 × Ar), 130.04 (Ar),
129.98 (2 × Ar), 129.88 (2 × Ar), 129.8 (2 × Ar), 129.7 (2 ×
Ar), 129.3 (Ar), 129.2 (Ar), 129.0 (Ar), 128.6 (Ar), 128.51 (2 ×
Ar), 128.48 (2 × Ar), 128.4 (2 × Ar), 128.2 (2 × Ar), 114.1 (2 ×
Ar), 107.2 (C-1′), 98.8 (C-1), 81.7 (C-4′), 76.3 (C-3′), 70.12 (C-
3), 70.06 (C-2), 69.3 (C-5), 66.7 (C-4), 64.5 (C-6), 64.4 (C-5′),
55.7 (OCH3), 55.3 (ArOCH3), 55.1 (C-2′), 36.2 (ArCH2S). HRMS
(ESI) calcd for (M + Na) C48H46O14S: 901.2501. Found: 901.2501.
2,3,5-Tri-O-acetyl-D-arabino-1,4-lactone (43). D-Arabinose (42,
15 g, 0.10 mmol) was dissolved in water (40 mL) and K2CO3 (17
g, 0.12 mmol) was added in portions. After the clear aqueous
solution was cooled to 0 °C, Br2 (6.0 mL, 0.11 mmol) was added
dropwise. The resulting orange solution was warmed to room
temperature and stirred for 18 h. The reaction mixture was quenched
by the addition of 88% formic acid (2.0 mL), and the solution was
concentrated. After addition of HOAc (20 mL), the reaction mixture
was concentrated again to remove any residual water. The yellowish
semisolid residue was redissolved in HOAc (20 mL) and cooled to
0 °C before Ac2O (90 mL) and concentrated H2SO4 (1 mL) were
added. After the reaction mixture was stirred at 50 °C for 18 h, it
was cooled to 0 °C and water (100 mL) was added. The aqueous
layer was extracted with CH2Cl2 (3 × 50 mL), and the combined
organic layer was washed with saturated aqueous NaHCO3 solution
(2 × 100 mL), dried over Na2SO4, filtered, and concentrated. The
crude product was purified by chromatography (4:1 hexane-EtOAc)
to afford 43 (20 g, 73%) as an oil. [R]D +24.6 (c, 0.6, CH2Cl2); 1H
NMR (600 MHz, CDCl3, δH) 5.57 (d, 1H, J2,3 ) 6.9 Hz, H-2),
5.46 (dd, 1H, J2,3 ) J3,4 ) 6.9 Hz, H-3), 4.57-4.53 (m, 1H, H-4),
4.67 (dd, 1H, J5a,5b ) 12.6 Hz, J4,5a ) 2.9 Hz, H-5a), 4.27 (dd, 1H,
J5a,5b ) 12.6 Hz, J4,5b ) 5.0 Hz, H-5b), 2.18 (s, 3H, CH3CO), 2.12
(s, 3H, CH3CO), 2.11 (s, 3H, CH3CO); 13C NMR (125 MHz, CDCl3,
δC) 170.2 (CdO), 169.8 (CdO), 169.4 (CdO), 168.2 (CdO), 77.4
(C-4), 72.6 (C-3), 72.2 (C-2), 62.1 (C-5), 20.6 (CH3CO), 20.5
1
to obtain 44 (1.51 g, 87%) as a white solid: H NMR (500 MHz,
CDCl3, δH) 5.62 (dd, 0.67H, J3,4 ) 9.2 Hz, J23 ) 2.1 Hz, H-3),
5.28 (d, 0.67H, J2,3 ) 2.3 Hz, H-2), 5.25 (ddd, 0.67H, J3,4 ) J4,5b
) 4.7 Hz, J4,5a ) 2.6 Hz, H-4), 5.21 (d, 0.33H, J2′,3′ ) 1.7 Hz,
H-2′), 5.05 (dd, 0.33H, J3′,4′ ) 4.7 Hz, J2′,3′ ) 1.7 Hz, H-3′),
4.42-4.32 (m, 0.67H, H-5a′ and H-4′), 4.30 (dd, 0.67H, J4,5a
)
2.6 Hz, J5a,5b ) 12.6 Hz, H-5a), 4.27-4.17 (m, 0.67H, H-5b′), 4.16
(dd, 0.67H, J4,5b ) 4.7 Hz, J5a,5b ) 12.6 Hz, H-5b), 2.15-2.05 (m,
12H, O(CO)CH3); 13C NMR (125 MHz, CDCl3, δC) 170.5-169.2
(6 × CdO), 167.5 (CdO ꢀ-Ac), 82.3 (C-4′), 80.5 (C-2′), 76.8 (C-
3′), 69.6 (C-2), 68.6 (C-3), 68.1 (C-4), 63.0 (C-5′), 61.6 (C-5),
21.0-20.3 (8 × CH3). HRMS (ESI) calculated for (M + Na)
C13H17DO9: 342.0906. Found: 342.0908.
Tolyl 2,3,5-Tri-O-acetyl-1-thio-r-D-[1-D]-arabinofuranoside
(45). To a solution of compound 44 (3.0 g, 9.40 mmol) and
p-thiocresol (1.4 g, 11.28 mmol) in CH2Cl2 (100 mL) at 0 °C was
added BF3 ·OEt2 (3.54 mL, 28.19 mmol) dropwise over 10 min.
The reaction mixture was stirred for 3 h at 0 °C, and then the
reaction was quenched by the addition of a saturated aqueous
NaHCO3 solution. The organic layer was separated, dried with
Na2SO4, filtered, and concentrated to a crude residue that was
purified by chromatography (4:1 hexane-EtOAc) to yield 45 (1.55
g, 43%) as a colorless oil: Rf (2:1 hexane-EtOAc); [R]D +149.2
1
(c, 0.3, CH2Cl2); H NMR (500 MHz, CDCl3, δH) 7.42-7.38 (m,
2H, Ar), 7.15-7.11 (m, 2H, Ar), 5.26 (d, 1H, J2,3 ) 2.2 Hz, H-2),
5.07 (dd, 1H, J3,4 ) 5.5 Hz, J2,3 ) 2.2 Hz, H-3), 4.47 (ddd, 1H, J3,4
) J4,5b ) 5.5 Hz, J4,5a ) 3.8 Hz, H-4), 4.39 (dd, 1H, J5a,5b ) 12.0
Hz, J4,5a ) 3.8 Hz, H-5a), 4.28 (dd, 1H, J5a,5b ) 12.0 Hz, J4,5b
)
5.5 Hz, H-5b), 2.33 (s, 3H, tolyl CH3), 2.12 (s, 3H, CH3CO), 2.10
(s, 3H, CH3CO), 2.09 (s, 3H, CH3CO); 13C NMR (125 MHz, CDCl3,
δC) 170.5 (CdO), 170.0 (CdO), 169.6 (CdO), 138.1 (Ar), 132.7
(2 × Ar), 129.8 (2 × Ar), 129.5 (Ar), 81.4 (C-2), 79.9 (C-4), 77.2
(C-3), 62.8 (C-5), 21.1 (tolyl CH3), 20.8 (CH3CO), 20.7 (2 ×
CH3CO). HRMS (ESI) calcd for (M + Na) C18H21DO7S: 406.1041.
Found: 406.1039.
Tolyl 1-Thio-r-D-[1-D]-arabinofuranoside (46). To a solution
of compound 45 (1.55 g, 4.04 mmol) in 1:1 CH2Cl2-CH3OH (60
mL) was added 1 M NaOCH3 in CH3OH (5 mL). After stirring for
16 h at room temperature, the reaction mixture was neutralized with
Amberlite IR-120 H+ resin, filtered, and concentrated. The resulting
oil was purified by chromatography (10:1, CH2Cl2-CH3OH) to
afford 46 (0.88 g, 85%) as a white solid: Rf 0.34 (10:1,
1
CH2Cl2-CH3OH); [R]D +208.0 (c, 0.2, CH2Cl2); H NMR (500
MHz, CD3OD, δH) 7.42-7.39 (m, 2H, Ar), 7.13-7.09 (m, 2H,
Ar), 4.00-3.89 (m, 3H, H-2, H-3, H-4), 3.75 (dd, 1H, J5a,5b ) 12.2
Hz, J4,5a ) 2.7 Hz, H-5a), 3.63 (dd, 1H, J5a,5b ) 12.2 Hz, J4,5b
)
4.6 Hz, H-5b), 2.30 (s, 3H, tolyl CH3); 13C NMR (125 MHz,
CD3OD, δC) 138.7 (Ar), 133.4 (2 × Ar), 132.4 (Ar), 130.6 (2 ×
Ar), 84.3 (C-4), 83.3 (C-2), 77.7 (C-3), 62.5 (C-5), 21.1 (tolyl CH3).
HRMS (ESI) calcd for (M + Na) C12H15DO4S: 280.0724. Found:
280.0726.
9
12946 J. AM. CHEM. SOC. VOL. 131, NO. 36, 2009

2,3-Anhydrosugar Migration-Glycosylation Mechanism
A R T I C L E S
Tolyl 2,3-Anhydro-5-O-benzoyl-1-thio-r-D-[1-D]-lyxofurano-
side (47). Compound 46 (0.88 g, 3.42 mmol), benzoic acid (0.63
g, 5.13 mmol), and triphenylphosphine (5.48 g, 10.26 mmol) were
dissolved in THF (50 mL) at 0 °C. Diisopropylazodicarboxylate
(2.09 mL, 10.6 mmol) was added dropwise over a period of 10
min. After complete addition of the reagent, the reaction mixture
was warmed to room temperature and was stirred for 45 min. The
solution was subsequently concentrated to a crude oil that was
purified by chromatography (8:1 hexane-EtOAc) to obtain 47 (0.82
g, 70%) as a white amorphous solid: Rf 0.81 (2:1 hexane-EtOAc);
NMR spectrum was recorded. Rf 0.49 (3:1 hexane-EtOAc); [R]D
-40.1 (c, 0.2, CH2Cl2); 1H NMR (500 MHz, CDCl3, δH) 7.38-7.27
(m, 7H, Ar), 7.14-7.11 (m, 2H, Ar), 5.15 (s, 0.5H, H-1), 4.80 (d,
1H, J ) 11.6 Hz, PhCH2), 4.63-4.59 (m, 1H, H-4), 4.51 (d, 1H,
J ) 11.6 Hz, PhCH2), 4.45 (dd, 1H, J5a,5b ) 11.9 Hz, J4,5a ) 4.4
Hz, H-5a), 4.30 (dd, 1H, J5a,5b ) 11.9 Hz, J4,5b ) 7.6 Hz, H-5b),
4.19 (dd, 1H, J3,OH ) 11.2 Hz, J3,4 ) 4.3 Hz, H-3), 3.76 (s, 1H,
H-2), 3.08 (d, 1H, J3,OH ) 11.2 Hz, OH), 2.34 (s, 3H, tolyl CH3),
2.11 (s, 1.5H, CH3CO); 13C NMR (125 MHz, CDCl3, δC) 170.9
(CdO), 137.9 (Ar), 136.7 (Ar), 131.7 (2 × Ar), 130.1 (2 × Ar),
129.4 (Ar), 128.6 (2 × Ar), 128.1 (3 × Ar), 106.2 (C-1), 81.1 (C-
4), 75.8 (C-3), 69.8 (PhCH2), 64.1 (C-5), 58.1 (C-2), 21.1 (tolyl
CH3), 21.0 (CH3CO). HRMS (ESI) calcd for (M + Na)+ C21H24O5S,
411.1237; C21H23DO5S, 412.1299; C21H21D3O5S, 414.1425;
C21H20D4O5S, 415.1488. Found: 411.1238, 412.1300, 414.1426,
415.1487, respectively.
1
[R]D +155.3 (c, 0.1, CH2Cl2); H NMR (500 MHz, CDCl3, δH)
8.10-8.06 (m, 2H, Ar), 7.60-7.55 (m, 1H, Ar), 7.48-7.41 (m,
4H, Ar), 7.15-7.12 (m, 2H, Ar), 4.56 (dd, 1H, J5a,5b ) 11.3 Hz,
J4,5a ) 5.7 Hz, H-5a), 4.50 (dd, 1H, J5a,5b ) 11.3 Hz, J4,5b ) 6.0
Hz, H-5b), 4.29 (dd, 1H, J4,5b ) 6.0 Hz, J4,5a ) 5.7 Hz, H-4), 3.94
(d, 1H, J2,3 ) 2.8 Hz, H-2), 3.84 (d, 1H, J2,3 ) 2.8 Hz, H-3), 2.35
(s, 3H, tolyl CH3); 13C NMR (125 MHz, CDCl3, δC) 166.2 (CdO),
138.4 (Ar), 133.4 (2 × Ar), 133.2 (2 × Ar), 129.9 (2 × Ar), 129.8
(2 × Ar), 128.6 (Ar), 128.4 (2 × Ar), 74.2 (C-4), 62.4 (C-5), 57.6
(C-2), 55.7 (C-3), 21.1 (tolyl CH3). HRMS (ESI) calcd for (M +
Na) C19H17DO4S: 366.0881. Found: 366.0878.
Methyl 5-O-(50% Acetyl-D3)-2-deoxy-2-p-thiotolyl-ꢀ-D-(50%
1-D)-xylofuranosyl)-(1f6)-2,3,4-tri-O-benzoyl-ꢀ-D-mannopyra-
noside (56). Thioglycoside 53 (a 1:1 mixture of 51 and 52, 81 mg,
0.29 mmol) and 4,4,5,5-tetramethyl-2-(1-naphthyl)-1,3-dioxolane
54 (37 mg, 0.14 mmol) were dissolved in CDCl3, and the 1H NMR
spectrum was recorded. After removal of CDCl3, the resulting
mixture was redissolved in CH2Cl2 (20 mL) with acceptor 30 (145
mg, 0.29 mmol) and 4 Å molecular sieves (250 mg). After the
mixture was stirred at room temperature for 30 min, Cu(OTf)2 (104
mg, 0.29 mmol) was added, and the mixture was stirred for 5 h.
The mixture was then filtered, the filtrate was concentrated, and
the 1H NMR spectrum was recorded. After neutralization with
triethylamine, the crude residue was then purified by chromatog-
raphy (2:1, hexane-EtOAc) to afford 56 and the 1H NMR spectrum
was recorded. Rf 0.24 (3:1 hexane-EtOAc); [R]D -70.5 (c, 0.3,
CH2Cl2); 1H NMR (500 MHz, CDCl3, δH) 8.12-8.08 (m, 2H, Ar),
7.98-7.94 (m, 2H, Ar), 7.83-7.80 (m, 2H, Ar), 7.61-7.57 (m,
1H, Ar), 7.54-7.50 (m, 1H, Ar), 7.47-7.33 (m, 7H, Ar), 7.27-7.24
Tolyl 2,3-Anhydro-1-thio-r-D-[1-D]-lyxofuranoside (48). To
a solution of 47 (1.02 g, 2.97 mmol) in 1:1 CH2Cl2-CH3OH (80
mL) was added 1 M NaOCH3 in CH3OH (4 mL). After stirring for
2 h at room temperature, the reaction mixture was neutralized with
Amberlite IR-120 H+ resin, filtered, and concentrated. The resulting
oil was purified by chromatography (4:1 hexane-EtOAc) to afford
48 (0.59 g, 83%) as a white solid: Rf 0.34 (2:1 hexane-EtOAc);
1
[R]D +128.7 (c, 0.2, CH2Cl2); H NMR (500 MHz, CDCl3, δH)
7.45-7.41 (m, 2H, Ar), 7.16-7.12 (m, 2H, Ar), 4.07 (ddd, 1H,
J4,5a ) J4,5b ) 5.0 Hz, J3,4 ) 0.6 Hz, H-4), 3.91 (d, 1H, J2,3 ) 2.8
Hz, H-2), 3.90-3.85 (m, 2H, H-5a, H-5b), 3.77 (dd, 1H, J2,3 ) 2.8
Hz, J3,4 ) 0.6 Hz, H-3), 2.34 (s, 3H, tolyl CH3), 1.88 (br, 1H, OH);
13C NMR (125 MHz, CDCl3, δC) 138.4 (Ar), 133.4 (2 × Ar), 129.9
(2 × Ar), 128.6 (Ar), 76.4 (C-4), 61.5 (C-5), 57.1 (C-2), 55.7 (C-
3), 21.1 (tolyl CH3). HRMS (ESI) calcd for (M + Na) C12H13DO3S:
262.0619. Found: 262.0619.
(m, 2H, Ar), 7.15-7.12 (m, 2H, Ar), 5.94 (dd, 1H, J4,5 ) J3,4
)
10.1 Hz, H-4), 5.85 (dd, 1H, J3,4 ) 10.1 Hz, J2,3 ) 3.3 Hz, H-3),
5.70 (dd, 1H, J2,3 ) 3.3 Hz, J1,2 ) 1.8 Hz, H-2), 5.15 (s, 0.5H,
H-1′), 4.95 (d, 1H, J1,2 ) 1.8 Hz, H-1), 4.65-4.60 (m, 1H, H-4′),
4.44-4.35 (m, 2H, H-5a′, H-5b′), 4.22-4.17 (m, 2H, H-5, H-3′),
4.08 (dd, 1H, J6a,6b ) 10.9 Hz, J5,6a ) 2.6 Hz, H-6a), 3.84 (s, 1H,
H-2′), 3.64 (dd, 1H, J6a,6b ) 10.9 Hz, J5,6b ) 3.7 Hz, H-6b), 3.54
(d, 1H, J3′,OH ) 11.9 Hz, OH), 3.51 (s, 3H, OCH3), 2.35 (s, 3H,
tolyl CH3), 2.06 (s, 1.5H, CH3CO); 13C NMR (125 MHz, CDCl3,
δC) 170.7 (CdO), 165.4 (3 × CdO), 137.7 (Ar), 133.44 (Ar),
133.42 (Ar), 133.1 (Ar), 131.5 (2 × Ar), 130.1 (2 × Ar), 130.0 (2
× Ar), 129.8 (2 × Ar), 129.7 (2 × Ar), 129.6 (Ar), 129.3 (Ar),
129.2 (Ar), 129.1 (Ar), 128.53 (2 × Ar), 128.47 (2 × Ar), 128.2
(2 × Ar), 106.4 (C-1′), 98.8 (C-1), 81.4 (C-4′), 75.8 (C-3′), 70.11
(C-2), 70.05 (C-3), 69.3 (C-5), 66.7 (C-4), 64.3 (C-5′), 64.1 (C-6),
58.1 (C-2′), 55.7 (OCH3), 21.1 (tolyl CH3), 20.9 (CH3CO). HRMS
(ESI) calcd for (M + Na)+ C42H42O13S, 809.2238; C42H41DO13S,
810.2301; C42H39D3O13S, 812.2427; C42H38D4O13S, 813.2489.
Found: 809.2237, 810.2305, 812.2430, 813.2493, respectively.
Tolyl 2,3-Anhydro-5-O-(acetyl-D3)-1-thio-r-D-(50%-1-D)-
lyxofuranoside (52). Thioglycoside 50 (a 1:1 mixture of 48 and
49, 315 mg, 1.32 mmol), DMAP (243 mg, 1.98 mmol), and
dicyclohexylcarbodiimide (546 mg, 2.65 mmol) were dissolved in
CH2Cl2 (20 mL) at room temperature. CD3CO2D (125 mg, 1.98
mmol) was added dropwise, and after complete addition, the
reaction mixture was stirred for 2 h. The solution was subsequently
filtered, and the filtrate was concentrated to a crude oil that was
purified by chromatography (5:1 hexane-EtOAc) to obtain 52 (301
mg, 80%). Rf 0.49 (2:1 hexane-EtOAc); [R]D +211.2 (c, 0.3,
CH2Cl2); 1H NMR (500 MHz, CDCl3, δH) 7.44-7.40 (m, 2H, Ar),
7.15-7.12 (m, 2H, Ar), 5.49 (s, 0.5H, H-1), 4.31 (dd, 1H, J5a,5b
)
11.4 Hz, J4,5a ) 5.9 Hz, H-5a), 4.23 (dd, 1H, J5a,5b ) 11.4 Hz, J4,5b
) 5.8 Hz, H-5b), 4.13 (ddd, 1H, J4,5a ) 5.9 Hz, J4,5b ) 5.8 Hz, J3,4
) 0.7 Hz, H-4), 3.91 (d, 1H, J2,3 ) 2.9 Hz, H-2), 3.75 (dd, 1H, J2,3
) 2.9 Hz, J3,4 ) 0.7 Hz, H-3), 2.34 (s, 3H, tolyl CH3); 13C NMR
(125 MHz, CDCl3, δC) 170.6 (CdO), 138.4 (Ar), 133.3 (2 × Ar),
129.9 (2 × Ar), 128.6 (Ar), 87.1 (C-1), 74.2 (C-4), 62.0 (C-5),
57.6 (C-2), 55.6 (C-3), 21.1 (tolyl CH3). HRMS (ESI) calcd for
(M + Na) C14H13D3O4S, 306.0850; C14H12D4O4S, 307.0913. Found:
306.0849, 307.0913.
NMR Experiments Carried Out To Probe Formation of
Episulfonium Ion Intermediates. An NMR tube containing a
solution of compound 1 (25 mg, 0.07 mmol) in CD2Cl2 (1 mL)
was cooled to -78 °C, and a spectrum was recorded at 400 MHz.
The tube was then removed from the spectrometer, and TMSOTf
(13 µL, 0.07 mmol) was quickly added at -78 °C. After returning
the sample to the spectrometer, the sample was warmed to room
temperature in 10° increments over 2 h, and a spectrum was
recorded at each temperature. A similar series of experiments was
carried out using 18 (25 mg, 0.07 mmol) in CD2Cl2 (1 mL) and
SnCl4 (8 µL, 0.07 mmol).
Calculation of Kinetic Isotope Effects. The procedure used was
analogous to that reported earlier by El-Badri et al.41 The 1H NMR
spectra of the mixture of donor 53 and internal standard 54, the
crude reaction mixture, and the product 2-thiotolyl-2-deoxy-ꢀ-
xylofuranoside (55 or 56) were recorded in CDCl3 at 500 MHz.
Benzyl 5-O-(50% Acetyl-D3)-2-deoxy-2-p-thiotolyl-ꢀ-D-(50%
1-D)-D-xylofuranoside (55). Thioglycoside 53 (a 1:1 mixture of
51 and 52, 52 mg, 0.19 mmol) and 4,4,5,5-tetramethyl-2-(1-
naphthyl)-1,3-dioxolane (54, 24 mg, 0.09 mmol) were dissolved
in CDCl3, and the 1H NMR spectrum was recorded. After removal
of CDCl3, the resulting mixture was redissolved in CH2Cl2 (20 mL)
with benzyl alcohol (29, 30 mg, 0.28 mmol) and 4 Å molecular
sieves (800 mg). The mixture was heated at reflux for 1.5 h before
1
it was filtered. The filtrate was concentrated, and the H NMR
spectrum was recorded. The crude residue was then purified by
1
chromatography (5:1 hexane-EtOAc) to afford 55, and the H
9
J. AM. CHEM. SOC. VOL. 131, NO. 36, 2009 12947

A R T I C L E S
Hou et al.
For each of spectra, the protons of the acetate group on O5, the
anomeric, and the internal standard protons were integrated on an
expanded region of the desired peak. The R-DKIE for each glycosy-
lation was determined using eq 144 (KIE ) ln(1 - F)/ln[1 - (FR/
R0)]), where F is the fractional conversion of donor 53 (the yield of
55 or 56) and R and R0 are the ratios of the anomeric protons
integrations in the product and the starting donor, respectively. These
ratios correspond to the D/H ratios in 55 (or 56) and 53.
and E3 ring conformers of 13a to investigate ring conformation.
The constrained geometry optimizations were conducted at the
B3LYP/6-31G** level of theory on structures 12a and 19 while
keeping the C1-S bond and the C2-S bond fixed at 1.901 and
1.871 Å, respectively. Single-point energies of these geometries
were subsequently calculated at the B3LYP/6-311++G** level.
Acknowledgment. This work was supported by the Natural
Sciences and Engineering Research Council of Canada and The
Alberta Ingenuity Centre for Carbohydrate Science. We thank
Professors Keith Woerpel (University of California, Irvine), Rik
Tykwinski (University of Alberta), and Andrew Bennet (Simon
Fraser University) for helpful discussions.
Gas-Phase ab Initio Geometry Optimizations. Ab initio
molecular orbital and density functional theory calculations were
performed using Gaussian 03.55 Episulfonium ions 19 and 12a and
oxocarbenium ions 20 and 13a were studied at the Hartree-Fock
(HF) and density functional theory (B3LYP) levels with the
6-31G** basis set. An initial geometry optimization was done on
each structure with no geometrical restraints. On each of the
resulting geometries, a single-point energy calculation was done
using the B3LYP/6-311++G** level of theory. Two additional
Supporting Information Available: Details on the synthesis
of compounds not described in the main text, 1H and 13C NMR
spectra for new compounds, and complete ref 55. This material
is available free of charge via the Internet at http://pubs.acs.org.
3
calculations at the same level of theory were also done on the E
(55) Frisch, M. J.; et al. Gaussian 03, version C.02, 2004; Gaussian, Inc.:
Wallingford, CT.
JA9029945
9
12948 J. AM. CHEM. SOC. VOL. 131, NO. 36, 2009