How the substituent at O-3 of N-acetylglucosamine impacts glycosylation at O-4: A comparative study

Article abstract of DOI:10.1021/jo901616p

(Chemical Equation Presented) An assessment of the relative reactivities of the 4-OH of N-acetylglucosamine acceptors bearing simple protecting groups, β-linked or α-linked D or L sugars at O-3 is presented, using a per-O-acetylated α-D-glucosyl trichloro

Full text of DOI:10.1021/jo901616p

pubs.acs.org/joc
How the Substituent at O-3 of N-Acetylglucosamine Impacts
Glycosylation at O-4: A Comparative Study
Jenifer L. Hendel, Jo-Wen Wang, Trudy A. Jackson, Karolyn Hardmeier,
Richelle De Los Santos, and France-Isabelle Auzanneau*
Department of Chemistry, University of Guelph, Guelph, Ontario, N1G 2W1 Canada
fauzanne@uoguelph.ca
Received July 24, 2009
An assessment of the relative reactivities of the 4-OH of N-acetylglucosamine acceptors bearing
simple protecting groups, β-linked or R-linked D or L sugars at O-3 is presented, using a per-O-
acetylated R-D-glucosyl trichloroacetimidate donor under activation by BF3 OEt2. The presence of
3
either an acyl or carbonate protecting group at O-3 did not impact the reactivity at O-4 with all
glycosylations proceeding successfully. On the other hand, the presence of peracetylated sugars at
O-3 of N-acetylglucosamine acceptors did impact the reactivity of the 4-OH. The acceptors with an
R-D-Man, β-D-Gal, or β-D-Glc at O-3 reacted promptly. In comparison, the acceptors bearing a β-L-
Fuc, R-L-Fuc, or R-L-Rha underwent glucosylation slowly, and unreacted acceptor was recovered
from the reaction mixtures. Systematic searches carried out on the disaccharide acceptors and
trisaccharide products carrying either a peracetylated β-D-Gal or β-L-Fuc at O-3 of the glucosamine
residue suggest that, for these two acceptors, a conformational reorientation necessary around the
fucosidic linkage contributes to the lower reactivity of the β-fucosylated acceptor. The acceptors
bearing a β-linked D-Gal, D-Glc, or L-Fuc residue at O-3 each gave trisaccharide products that were
mostly stable in the reaction conditions. In contrast, the R-linked residues at O-3 were rather unstable
in these reaction conditions and the degradation of either the acceptors or trisaccharide products led
to low glycosylation yields. In these later reactions, it was impossible to clearly assess which of the
acceptor or product underwent degradation as comigration and detection issues prevented us from
following these glycosylations by TLC or RP-HPLC. In contrast, the glycosylation of an acceptor
carrying an R-linked perbenzylated L-Fuc residue at O-3 could be easily monitored by RP-HPLC.
The data obtained when monitoring this glycosylation showed that the acceptor underwent prompt
glycosylation but a decrease in the absorbance peak corresponding to the trisaccharide along with the
appearance of a peak corresponding to a perbenzylated fucose hemiacetal indicated that the
trisaccharide product was unstable in the reaction conditions.
Introduction
reactions.1-3 This lack of reactivity often makes the synthesis
of biologically important oligosaccharides containing N-
acetylglucosamine residues glycosylated at O-4 an arduous
task. The low reactivity exhibited by this hydroxyl group has
been attributed to steric hindrance around the O-4 position1
as well to the formation of a hydrogen-bonding network
involving the amide hydrogen.2 In addition, poor yields also
often result from the formation of stable glycosyl imidate
It is well-known that the hydroxyl group at C-4 of N-
acetylglucosamine is a poor nucleophile and has reduced
reactivity when compared to other acceptors in glycosylation
(1) Paulsen, H. Angew. Chem., Int. Ed. Engl. 1982, 21, 155–173.
(2) Crich, D.; Dudkin, V. J. Am. Chem. Soc. 2001, 123, 6819–6825.
(3) Liao, L; Auzanneau, F.-I. Org. Lett. 2003, 5, 2607–2610.
DOI: 10.1021/jo901616p
r
Published on Web 10/07/2009
J. Org. Chem. 2009, 74, 8321–8331 8321
2009 American Chemical Society

Hendel et al.
JOCArticle
side products.3 The most common method to overcome
both the low reactivity at this position and the formation
of glycosyl imidates is to replace the N-acetyl group with,
for instance, an N-phthalimido (N-Phth), N-trichloro-
ethoxycarbonyl (N-Troc), or azido substituent. More
recently Crich and Vinod have also reported that the
simultaneous protection of the N-acetyl group and O-3
with an oxazolidone allowed for efficient glycosylation to
occur at O-4 of N-acetylglucosamine.4 Although rare,
there are reports in the literature of successful glycosyla-
tions at O-4 of N-acetylglucosamine glycosyl acceptors.5
In fact, we have recently described reaction conditions6
that allow such successful glycosylations to take place.
Glucosylation at O-4 of an N-acetylglucosamine mono-
saccharide acceptor, using the R-trichloroacetimidate of
peracetylated glucopyranose as a donor, proceeded with-
out the concomitant formation of glycosyl imidates and in
more than 90% yield when carried out under activation
TABLE 1. Glucosylation of Monosaccharide Acceptors a
entry
acceptor
product
yield (%)
1
2
3
4
2
3
4
5
6
7
8
9
88
76
87
90
aReagents and conditions: BF3 OEt2 (2 equiv), 1 (5 equiv), CH2Cl2,
3
40 °C, 1 h
SCHEME 1. Glycosylation of Monosaccharide Acceptors
with excess BF3 OEt2 (2 equiv) at room temperature. We
3
have hypothesized6 that in these conditions 1 equiv of
BF3 OEt2 interacted noncovalently with the nucleophilic
3
N-acetyl group and that the second one promoted glyco-
sylation. Therefore, we concluded that as long as donor
and acceptor could withstand activation with multiple
equivalents of BF3 OEt2 at elevated temperatures, glyco-
3
sylations at O-4 of a glucosamine acceptor with a trichloro-
acetimidate donor may proceed in synthetically useful
yields. However, the result described above constitutes
only one successful example of such glycosylation and
there is limited literature to clarify when glycosylation at
O-4 of N-acetylglucosamine is feasible. In this paper, we
report our results when attempting to glucosylate at O-4 of
N-acetylglucosamine in various mono- and disaccharide
acceptors using the reaction conditions mentioned above.
We have examined the impact that the substituent at O-3 of
N-acetylglucosamine acceptors has on the reactivity of the
neighboring 4-OH by attempting glycosylations of accep-
tors bearing various protecting groups or sugar residues at
O-3. The disaccharide acceptors include both D and L series
sugars, linked at O-3 through either R or β glycosidic
bonds. Thus, we report that while simple acyl groups at
O-3 allow for such reactions to proceed in high yield,
having a sugar residue at O-3 may impact negatively and
drastically the outcome of the glycosylation. We show that
yields are sometimes reduced due to a lower reactivity at
O-4 but also in some cases due to the fast degradation of
product in these rather harsh conditions. Indeed, trends
indicate that the efficiency of the glycosylation at O-4 is
affected by the structure of the sugar residue at O-3 (D vs. L)
and that the stability of the trisaccharide formed is affected
by the stereochemistry of the glycosidic linkage at O-3
(R vs. β). Finally, we describe the use of RP-HPLC to
monitor glycosylation of an acceptor bearing a perbenzy-
lated residue at O-3.
Results and Discussion
In this study the known7 trichloroacetimidate 1 was used
to attempt glycosylation of four monosaccharide acceptors
(2-5), three disaccharide acceptors that carried a β-linked
glycoside at O-3 (10-12), and four disaccharide acceptors
that carried an R-linked glycoside at O-3 (17-19, 23). While
acceptors 2,6 3,8 10,9 19,10 and 2311 are known, the prepara-
tion of monosaccharide acceptors 4 and 5 as well as that of
disaccharide acceptors 11, 12, 17, and 18 is described in the
Supporting Information for this paper.
Glycosylation of Monosaccharide Acceptors (Scheme 1,
Table 1). We have reported6 that under activation with 2
equiv of BF3 OEt2, the R-trichloroacetimidate 1 (5 equiv)
3
could be easily coupled to the 3-chloroacetylated acceptor 2.
Indeed whether the reaction was left to proceed at room
temperature for 1 h or stirred at 40 °C for 10 min, the
disaccharide 6 has been isolated in about 90% yield.6 In this
study, all the glycosylations were first attempted by using 5
equiv of donor 1 and 2 equiv of BF3 OEt2 and were left to
3
proceed for 1 h at 40 °C. As expected, these conditions led to
an excellent yield (88%) of disaccharide 6 when coupling
donor 1 with chloroacetate acceptor 2 (Table 1, entry 1). In
fact the results presented in Table 1 show that each of the
coupling reactions involving donor 1 and the monosacchar-
ide acceptors (2-5) proceeded successfully. There was essen-
tially no difference between the relative reactivities of the
acceptors bearing an Alloc, or Troc (4 and 5, entries 3 and 4)
(4) Crich, D.; Vinod, A. U. Org. Lett. 2003, 5, 1297–1300. Crich, D.;
Vinod, A. U. J. Org. Chem. 2005, 70, 1291–1296.
(7) Schmidt, R. R.; Michel, J. Angew. Chem. 1980, 19, 731–732.
(8) Zhang, P.; Appleton, J.; Ling, C.-C.; Bundle, D. R. Can. J. Chem.
2002, 80, 1141–1161.
(9) Khare, D. P.; Hindsgaul, O.; Lemieux, R. U. Carbohydr. Res. 1985,
136, 285–308.
(10) Auzanneau, F.-I.; Hanna, H. R.; Bundle, D. R. Carbohydr. Res.
1993240, 16181
(11) Sakagami, M.; Hamana, H. Tetrahedron Lett. 2000, 41, 5547–5551.
(5) For example: Unverzagt, C.; Kunst, H. J. Prakt. Chem. 1992, 334,
570–578. Kiyoi, T.; Nakai, Y.; Kondo, H.; Ishida, H.; Kiso, M.; Hasegawa,
A. Bioorg. Med. Chem. 1996, 4, 1167–1176. Misra, A.; Ding, Y.; Lowe, J. B.;
Hindsgaul, O. Bioorg. Med. Chem. Lett. 2000, 10, 1505–1509. Holger, H.;
Kunz, H. Carbohydr. Res. 2007, 342, 541–557.
(6) Hendel, J. L.; Cheng, A.; Auzanneau, F.-I. Carbohydr. Res. 2008, 343,
2914–2923.
8322 J. Org. Chem. Vol. 74, No. 21, 2009

Hendel et al.
JOCArticle
SCHEME 2. Glucosylation of Acceptors Bearing a β-Linked Sugar at O-3
degradation when left in these conditions (2 equiv of
BF3 OEt2) for such an extended time.
TABLE 2. Glucosylation of Acceptors Bearing a β-Linked Sugar at
O-3a
3
yields (%)
Surprisingly, the glucosylation of acceptor 12 that carries a
β-L fucosyl residue at O-3 proceeded with much greater
difficulty than that of acceptors 10 and 11 that both carry
a β-D residue at O-3. While an analytical sample of the pure
trisaccharide 15 was isolated after multiple chromato-
graphies, comigration of acceptor 12 and trisaccharide 15
on normal phase silica gel precluded the monitoring of these
reactions by TLC. Thus, the two compounds were isolated
together by flash chromatography and 1H NMR was used to
determine the yields reported in Table 2 (entries 6 and 7).
Thus, after 1 h at 40 °C, the glucosylation of acceptor 12
(Table 2, entry 6) gave a 7:3 mixture of trisaccharide 15
(63%) and unreacted acceptor 12 (24%) supporting that the
β-fucosylated acceptor 12 was less reactive than acceptors 10
and 11 carrying respectively a β-linked D galactose or glucose
at O-3. Identical results within experimental error were
obtained when the glucosylation of β-fucosylated acceptor
12 was left to proceed at room temperature for 4 h (Table 2,
entry 7). While in all cases the trisaccharides were formed in
acceptable yields, these results suggest that the glycosyla-
tion at O-4 of N-acetylglucosamine acceptors that carry a
β-linked D sugar residue at O-3 is easier than that of acceptors
that carry a β-linked L unit at O-3.
entry acceptor temp (°C) time (h) product product acceptor
1
2
3
4
5
6
7
10
11
10
11
10
12
12
40
40
rt
1
1
1
1
18
1
13
14
13
14
13
15
15
93
88
88
91
rt
40
40
rt
61
63b
24b
18b
4
63b
aReagents and conditions: BF3 OEt2 (2 equiv), 1 (5 equiv), CH2Cl2.
3
bAcceptor and product were isolated as a mixture, the yield was
determined by NMR.
when compared to that of the chloroacetate 2, and the three
couplings gave the disaccharides 8, 9, and 6 in yields of
around 90%. In contrast, the acceptor 3 bearing a benzoyl
group at O-3 gave a significantly lower but still acceptable
76% yield of the disaccharide 7 (Table 1, entry 2), possibly
due to either electronic or steric effects of the phenyl ring.
Globally, acceptors carrying an acyl or carbonate group at
O-3 of N-acetylglucosamine reacted very well toward glyco-
sylation at O-4 with donor 1 activated with 2 equiv of
BF3 OEt2 at 40 °C.
3
Glycosylation of Disaccharide Acceptors Carrying a β-
Linked Residue at O-3 (Scheme 2, Table 2). We then focused
our attention on the coupling of the glucosyl R-trichloro-
acetimidate 1 with O-4 of N-acetylglucosamine acceptors
that carried a β-linked D or L sugar residue at O-3 (10-12).
These glycosylation reactions were first attempted by
using the same conditions as described above (5 equiv of 1,
A Possible Explanation to the Slower Glycosylation of the
β-L-Fucosylated Acceptor 12 vs. that of the β-D-Galactosy-
lated Acceptor 10. As with all chemical processes, the reacti-
vity of substrates in glycosylation reactions is largely con-
trolled by electronic and steric effects leading to the relative
stabilization of starting materials, transition states, inter-
mediates, and products. While quantum mechanical calcula-
tions are best suited to investigate the detailed electronic
behaviors that led to the observed differences in reacti-
vity between, for example, the β-D-galactosylated
acceptor 10 and the β-L-fucosylated acceptor 12, the size of
these molecules precludes the use of such high-level calcula-
tions. One may suggest that the two disaccharide acceptors
10 and 12 are quite similar in their electronic behaviors since
they both display a β-linked galactose-type sugar residue
at O-3. Indeed, electronically these acceptors only differ
by the presence of an additional acetyl substituent at C-6
of this galactose residue in acceptor 10 while acceptor 12 is
2 equiv of BF3 OEt2, 1 h at 40 °C). As can be seen in Table 2
3
(entries 1 and 2), the glucosylation of acceptors 10 and 11
that carry β-D-galactosyl and glucosyl residues, respectively,
at O-3 gave the corresponding trisaccharides 13 and 14 in
approximately 90% yield.
Even at room temperature (Table 2, entries 3 and 4),
acceptors 10 and 11 were promptly glycosylated, and after
1 h of reaction the trisaccharides 13 and 14 were isolated in
∼90% yield. Interestingly, when the coupling of 1 and 10 was
left to proceed at 40 °C for 18 h, the yield of trisaccharide 13
dropped to 61% suggesting that it was slowly undergoing
J. Org. Chem. Vol. 74, No. 21, 2009 8323

Hendel et al.
JOCArticle
deoxygenated at this position. In contrast, while the galac-
tose residue in acceptor 10 belongs to the D hexoses family
TABLE 3. Glycosidic Torsions in Disaccharides 10, 12, 16 and
Trisaccharides 13, 15 Global Minima
torsion anglesa
4
and thus adopts a C1 conformation, the fucose residue in
acceptor 12 belongs to the L hexoses family and thus adopts a
1C4 conformation. Therefore, it is reasonable to hypothesize
that conformational differences may contribute to the dif-
ference in reactivity toward glycosylation that was observed
between the β-D-galactosylated acceptor 10 and the β-L-
fucosylated acceptor 12. Indeed, differences in the confor-
mations of the acceptors may make O-4 less accessible in
acceptor 12 than in acceptor 10 or differences in the con-
formations of the products may make trisaccharide 15 less
stable and thus more difficult to obtain than trisaccharide 13.
It is generally accepted that branched trisaccharides are
much less flexible than the corresponding individual disac-
charides. Thus, we wondered if the lowest energy conforma-
tions that are favored in the more flexible disaccharide
analogues 10, 12, or 16 were still accessible in the branched
analogues 13 and 15, or if any glycosidic torsion could no
longer achieve the most favored orientation identified in the
simple disaccharides. To test this hypothesis, we carried out
systematic searches with Sybyl8.0 on compounds 10, 12, 13,
and 15 and as well as on disaccharide 16 in implicit CH2Cl2
and using the Tripos12 force field with the inclusion of the
PIM parameters13 for carbohydrates. The acetate
entry
compd
Φ1
Ψ1
Φ2
Ψ2
1
2
3
4
5
10
12
16
13
15
-72
80
96
-48
-64
-66
-72
-107
-130
-109
-73
142
119
145
aΦ1, Ψ1 = β-Gal/Fuc-(1f3)-GlcNAc; Φ, Ψ2 = β-Glc-(1f4)-
GlcNAc.
linkage in trisaccharides 13 and 15. The Φ/Ψ maps for this
glycosidic linkage in disaccharide 16 as well as in trisacchar-
ides 13 and 15 are given in the Supporting Information and
were essentially identical with one another with only an
apparent restricted mobility around this β-D-Glc-(1f4)-
GlcNAc bond in the trisaccharides. As seen in Table 3, the
Φ/Ψ torsions found for the global minimum A identified for
disaccharide 16 (entry 3) were essentially conserved in the
global minima found for trisaccharides 13 (Φ2/Ψ2, entry 4)
and 15 (Φ2/Ψ2, entry 5). Therefore, we concluded that the
presence at O-3 of the galactose or fucose residues in the
disaccharide acceptors 10 and 12 did not prevent the glucose
residue introduced at O-4 of the N-acetylglucosamine from
adopting the most favored glycosidic torsions found in
disaccharide 16.
C-O-C-C torsions were rotated in 180° increments and
14
ꢀ
~
based on the study by Gonzalez-Outeirino et al. the
secondary acetate H-C-O-C torsions were rotated in 30°
increments from -60° to þ60° while the same torsions in
primary acetates were rotated in 30° increments from 0 to
360°. With use of the known15 preferred orientations of the
acetamido group in N-acetylglucosamine, the H-C-N-C
and C-N-C-C torsions were rotated in 180° increments.
The hydroxymethyl and benzyl groups rotating bonds were
sampled with 120° increments reflecting their expected stag-
gered preferences and the O-H group was rotated in 120°
increments to include both H-bonding and non-H-bonding
orientations. Most of the computational effort was
expended on rotations around the glycosidic linkages as
the Φ and Ψ torsions were rotated in 10° increments. Only
Next, we assessed the impact that the presence of a glucose
residue at O-4 of GlcNAc had on the glycosidic torsions of
the O-3 galactose or fucose residues. Panels a and b of
Figure 1 show the Φ/Ψ maps for the β-D-galactosyl- and
β-L-fucosyl-(1f3)-GlcNAc glycosidic bonds for the disac-
charides 10 and 12, respectively, while panels c and d of
Figure 1 show the Φ1/Ψ1 maps for the same glycosidic
linkages in trisaccharide 13 and trisaccharide 15, respectively.
It is evident from Figure 1a,b that the β-D-galactosylated
acceptor 10 is much less flexible than the β-L-fucosylated
acceptor 12. Indeed while within 3 kcal mol-1 of the global
3
minimum only one local minimum (Figure 1a, B) was found
for acceptor 10, three such local minima (Figure 1b, B-D)
were identified for the fucosylated analogue 12. As could be
expected, the two acceptors 10 and 12 adopt global minima
(Table 3, entries 1 and 2; Figure 1a,b) that favor the (-) or (þ)
exoanomeric orientation, respectively, and in both cases allow
the formation of a hydrogen bond between the galactose or
fucose ring oxygen and the GlcNAc 4-OH.
the conformations found within 10 kcal mol-1 of the lowest
3
energy conformation were retained and clustered based on
the Φ and Ψ torsions.
When comparing the Φ/Ψ map obtained for the β-D-
Gal-(1f3)-GlcNAc glycosidic bond in acceptor 10
(Figure 1a) with the Φ1/Ψ1 map obtained for the same bond
in trisaccharide 13 (Figure 1c) one observes a considerable
loss of flexibility around this bond with the local minimum B
found for the disaccharide no longer found in the branched
trisaccharide. However, as can be seen in these maps as well
as in Table 3 (compare entries 1 and 4) and Figure 2a, the
global minimum (A) adopted by this linkage in disaccharide
10 (Figure 2a, gray structure) is essentially conserved in
trisaccharide 13 (Figure 2a, black structure) with only a
rotation of less than 30° around Ψ1 to accommodate more
comfortably the glucose residue introduced at the vicinal
O-4. In sharp contrast, a very different picture emerges when
comparing the Φ/Ψ map obtained for the β-L-Fuc-(1f3)-
GlcNAc glycosidic bond in acceptor 12 (Figure 1b) with the
The lowest energy representatives of each family were
further minimized and the torsion angles measured for the
final global minima identified for the disaccharides and
trisaccharides are listed in Table 3. We first compared the
conformational preference of the glycosidic β-D-Glc-(1f4)-
GlcNAc bond in the protected disaccharide 16 to the same
(12) SYBYL 8.0; Tripos Associates, 1699 S. Hanley Road, Suite 303, St
Louis, MO 63144.
ꢀ
(13) Imberty, A.; Hardman, K. D.; Carver, J. P.; Perez, S. Glycobiology
1991, 1, 631–642.
ꢀ
~
(14) Gonzalez-Outeirino, J.; Nasser, R.; Anderson, J. E. J. Org. Chem.
2005, 70, 2486–2493.
(15) Flower, P.; Bernet, B.; Vasella, A. Helv. Chim. Acta 1996, 79, 269–
287.
8324 J. Org. Chem. Vol. 74, No. 21, 2009

Hendel et al.
JOCArticle
FIGURE 1. Systematic searches Φ/Ψ maps for the 1f3 glycosidic bonds: (a) β-D-Gal-(1f3)-GlcNAc in disaccharide 10; (b) β-L-Fuc-(1f3)-
GlcNAc in disaccharide 12; (c) Φ1/Ψ1 β-D-Gal-(1f3)-GlcNAc in trisaccharide 13; and (d) Φ/Ψ1 β-L-Fuc-(1f3)-GlcNAc in trisaccharide 15.
Minimized global minima are identified as A; local minima found within 3 kcal mol-1 of A are labeled B-D where appropriate.
3
become very rigid when compared to its flexibility in di-
saccharide 12 (Figure 1b). Most importantly, we observe that
the lowest energy conformation adopted around this glyco-
sidic bond in trisaccharide 15 (Figure 1d, A; Table 3, entry 5)
is very different from the lowest energy conformation found
for disaccharide 12 (Figure 1b, A; Table 3, entry 2).
Thus, it implies that either a þ190° or a -170° rotation
around the Ψ torsion had to occur during the glucosylation
at O-4. Indeed such a rotation is required since, as seen in
Figure 2b, if the fucosyl residue was to maintain the con-
formation that is favored in the disaccharide (gray) around
the glycosidic bond, its C-60 would collide with C-500 of the
glucose residue (black) introduced at O-4. Due to this atom
overlap, it is reasonable to expect this conformational shift to
occur prior to or concurrently to the glycosylation and as
such to affect the observed rate of the reaction. We subse-
quently assessed the energy difference between the global
minimum adopted by disaccharide 12 and the conformation
that it needs to assume to give trisaccharide 15. Thus,
acceptor 12 was minimized around the side chains and
protecting groups while the Φ/Ψ torsions of the β-L-
Fuc-(1f3)-GlcNAc bond was set and maintained at the
FIGURE 2. Molecular overlays of disaccharide and corresponding
values found for this glycosidic bond in the global minimum
trisaccharide global minima: (a) disaccharide acceptor 10 (gray) and
identified for trisaccharide 15 (Table 3 entry 5, Φ1/Ψ1).
trisaccharide 13 (black) and (b) disaccharide acceptor 12 (gray) and
Indeed, the resulting minimized conformation was found
trisaccharide 15 (black).
to be over 12 kcal mol-1 higher in energy than the global
minimum found for acceptor 12.
3
Φ1/Ψ1 map obtained for the same bond in trisaccharide 15
(Figure 1d). Indeed, the Φ1/Ψ1 map presented in Figure 1d
shows that this glycosidic bond in trisaccharide 15 has
The vicinal coupling constants measured for the sugar
units in compounds 10, 12, 13, and 15 supported an average
J. Org. Chem. Vol. 74, No. 21, 2009 8325

Hendel et al.
JOCArticle
SCHEME 3. Glucosylation of Acceptors Bearing an R-Linked Sugar at O-3
TABLE 4. Glucosylation of Acceptors Bearing an R-Linked Sugar at
However, these trisaccharides comigrated on silica gel with
either the acceptor or degraded glucosyl donor and the
reactions could not be followed by TLC.
O-3a
yield (%)
The yields reported in Table 4 were therefore estimated by
1H NMR on the mixtures obtained after chromatography.
Starting with the conditions established above (5 equiv of 1, 2
entry acceptor temp (°C) time (h) product product acceptor
1
2
3
4
5
6
7
8
9
17
17
17
17
17
18
18
18
19
19
19
40
40
40
rt
1
2
18
2
18
1
3
8
1
2
18
20
20
20
20
21
21
21
21
22
22
22
55b
41b
c
equiv of BF3 OEt2, 1 h at 40 °C), glycosylation of the D
-
3
54b
39b
26d
37d
mannosylated acceptor 17 gave trisaccharide 20 (Table 4,
entry 1) in 55% yield while no acceptor was recovered.
Extending the reaction time to 2 h (entry 2) gave a decreased
yield (41%) of the trisaccharide 20 while after 18 h (entry 3),
TLC analysis of the reaction mixture showed only trace
amounts of trisaccharide product 20. Since no acceptor 17
was recovered in any of these reactions, we concluded that it
was either degraded before being glycosylated or promptly
glucosylated to give the trisaccharide 20 that, in turn, under-
went fast degradation. Running the same glycosylation at
room temperature (entry 4 and 5) appeared to slow the
degradation of the trisaccharide 20, which was obtained in
54% or 39% yield after 2 or 18 h, respectively. Thus, it
appears that similarly to acceptors 10 and 11 that also carry D
sugar residues at O-3, the mannosylated acceptor 17 was
promptly glucosylated either at 40 °C or at room tempera-
ture. However, while the acceptors (10, 11) and trisacchar-
ides (13, 14) that carry β-D residues at O-3 were fairly stable
in these glycosylation conditions (Table 2), the starting
acceptor 17 or trisaccharide 20 that carry an R-D-mannosyl
residue at O-3 underwent fast degradation.
rt
40
40
40
40
40
40
34d
38d
c
c
-
-
27d
34d
43d
40d
10
11
c
-
c
-
aReagents and conditions: BF3 OEt2 (2 equiv), 1 (5 equiv), CH2Cl2.
3
bContaminated with some degraded donor, yield estimated by NMR.
cTrace amounts observed by TLC. dMixture of donor and acceptor,
yields estimated by NMR.
4C1 conformation for the glucose, galactose, and N-acetyl-
glucosamine rings and a 1C4 conformation of the fucose ring.
Unfortunately, signal overlap in 2D NOESY experiments at
300 K in deuterated chloroform prevented the acquisition of
any other usable experimental data to support the modeling
results described above. A more detailed study involving the
modeling of the reaction trajectory and transition states
would be required to fully understand these behaviors.
However, the results presented above suggest that the degree
of conformational reorientation necessary around the fuco-
sidic linkage when going from acceptor 12 to trisaccharide
15, as well as the reduction in conformational space acces-
sible around this bond, contributes to the slower glycosyla-
tion of the β-L-fucosylated acceptor 12 when compared to
that of the β-D-galactosylated acceptor 10.
Glycosylation of Disaccharide Acceptors Bearing an
r-Linked Residue at O-3 (Scheme 3, Table 4). In turn, we
focused our attention on the coupling of the glucosyl
R-trichloroacetimidate 1 with O-4 of N-acetylglucosamine
acceptors that carried peracetylated R-linked D or L sugar
residues at O-3 (17-19, Table 4).
Analytical samples of trisaccharides 20-22 were obtained
after multiple chromatographies and their structures
were confirmed by NMR and HR-ESI mass spectrometry.
Glucosylations of acceptors 18 and 19 that carry respec-
tively an R-L-fucosyl and R-L-rhamnosyl residue at O-3 were
then investigated (Table 4, entries 6-11). After 1 h at 40 °C
only 26% and 27% of fucosylated trisaccharide 21 or
rhamnosylated trisaccharide 22 were obtained, respectively,
while unreacted acceptors were recovered in 34% yield for
acceptor 18 and 43% for acceptor 19 (entries 6 and 9).
Allowing the glycosylations to proceed for longer reaction
times at 40 °C increased the yield of the trisaccharides 21 and
22 to 37% and 34%, respectively, but large amounts of
unreacted acceptors were still present (entries 7 and 10).
Finally, allowing the glucosylation of the acceptors 18 and 19
to run for 8 or 18 h, respectively, resulted in only trace
quantities of both of the acceptors and products observed by
TLC (entries 8 and 11). These results suggest that the O-4
8326 J. Org. Chem. Vol. 74, No. 21, 2009

Hendel et al.
JOCArticle
FIGURE 3. HPLC trace at 50 min reaction time for the glucosylation of acceptor 23 (2 equiv of BF3 OEt2 at 40 °C).
3
SCHEME 4
GlcNAc acceptors 18 and 19 carrying an R-L- sugar residue
at O-3 were less reactive than acceptor 17 that carries an R-D-
mannosyl residue at the same position. However, it is clear
from those reactions left to proceed at 40 °C for extended
reaction times that compounds (acceptors or trisaccharides)
carrying an R-linked sugar residue at O-3 of the GlcNAc unit
were prone to degradation (Table 4, entries 3, 8, and 11). In
contrast, O-4 glucosylation of the analogues bearing
β-linked sugar residue at O-3 led to considerably more stable
trisaccharides with, for example, 61% of the 3-O-β-D-galacto-
sylated trisaccharide 13 still isolated when the reaction was
left at 40 °C for 18 h (Table 2, entry 5).
Our conclusion that acceptors and trisaccharides bearing
an R-linked residue at O-3 of the GlcNAc unit could undergo
degradation was further supported by our study on the
glycosylation of the disaccharide acceptor 23 that contains
an R-linked perbenzylated fucose residue at O-3 of the
N-acetylglucosamine unit (Scheme 4). In contrast to the
glycosylations studied above, this reaction could be moni-
tored by HPLC on a C18 reverse phase column, using a
gradient solvent system starting at 65:35 CH3CN:H2O and
increasing to 90:10 CH3CN:H2O. The disappearance of the
acceptor and the formation of the desired trisaccharide as
well as the formation of side products was monitored with
use of UV detection at 220 nm. As can be seen in Figure 3, the
HPLC trace gave well-separated absorbance peaks for the
acceptor 23 and two newly formed compounds 24 and 25. In
contrast, the unreacted glycosyl donor and its derivatives
were observed close to the solvent front and with a low
absorbance since they do not carry aromatic substituents.
A semipreparative RP-HPLC column allowed the separa-
tion of compounds 23, 24, and 25. Compound 24 was
identified as the desired known6 trisaccharide while 25 was
identified as the known16 2,3,4-O-benzyl-R-L-fucopyranose.
Since acceptor 23, trisaccharide 24, and fucose 25 have
similar absorbance at 220 nm, the areas measured for these
three peaks in the HPLC trace could be correlated to the
relative amounts of the three products present at a given time
in the reaction mixture. Indeed, glycosylation of acceptor 23
with donor 1 could be followed directly by analytical RP-
HPLC: at each time point a 5 μL aliquot of the reaction
mixture was diluted in CH3CN (100 μL) containing NEt3
and 10 μL of this solution was injected in the HPLC. Three
reactions were followed by HPLC and the relative abun-
dance of 23, 24, and 25 as a function of time for each reaction
is shown in Figure 4.
The first reaction was carried out in the standard condi-
tions used above (5 equiv of donor 1, 2 equiv of BF3 Et2O at
3
40 °C, Figure 4a). In these conditions, after 5 min of reaction
it was observed that already 61% of trisaccharide 24 had
been formed while 39% of the acceptor remained. After 50
min of reaction the amount of trisaccharide 24 had slightly
increased to 64% while only 30% of the acceptor 23
remained and 10% of the tribenzylated fucose 25 had
formed. After 100 min of reaction, HPLC analysis of the
reaction showed complete disappearance of the formed
trisaccharide 24, 10% of residual starting acceptor 23, and
90% of the tribenzylated fucose 25. Interestingly, column
chromatography of the reaction mixture led to isolation of
(16) Dejter-Juszynski, M.; Flowers, H. M. Carbohydr. Res. 1971, 18, 219–
226.
J. Org. Chem. Vol. 74, No. 21, 2009 8327

Hendel et al.
JOCArticle
perbenzylated) fucosyl residue at O-3 of the GlcNAc accep-
tor (Table 4, entries 6 and 7). It is also clear that the
glucosylation of acceptor 23 occurred faster than its degra-
dation but that the formed trisaccharide 24 was unstable in
the presence of 2 equiv of BF3 Et2O and lost the fucosyl
3
residue. We also attempted coupling acceptor 23 and donor 1
at room temperature using only 1 equiv of BF3 Et2O
3
(Figure 4c). In these conditions, only 30% the trisaccharide
24 was formed within 5 min while 70% of the acceptor
remained unreacted. This ratio remained stable for 100 min
and no further glucosylation of 23 nor degradation of 23 and
24 was observed. These results suggest that glucosylation of
acceptor 23 by donor 1 requires the use of 2 equiv of
BF3 Et2O to proceed, and that in these conditions, it pro-
3
ceeds quickly at either room temperature or 40 °C. However,
we also observe that in these conditions, the trisaccharide
product 24 undergoes degradation by losing the R-fucosyl
residue at O-3 of GlcNAc. Indeed, such degradation has also
been observed by Shimizu et al.17 upon galactosylation of a
dichlorophtalimidoglucosamine O-4 acceptor carrying a
perbenzylated R-fucosylated unit at O-3.
Conclusion
It is evident from the studies above that the substituent at
O-3 of an N-acetylglucosamine acceptor impacts the reactiv-
ity of the neighboring 4-OH. An acyl or carbonate protecting
group at O-3 of the N-acetylglucosamine acceptor appears to
have little impact on the reactivity of the 4-OH. In contrast,
glycosylations of acceptors bearing various sugar residues at
O-3 showed that the configuration and nature of the sugar at
O-3 had a large impact on the reactivity of the 4-OH toward
glycosylation. Two trends seem to be emerging from the
results described above for the glycosylations of the accep-
tors bearing peracetylated β-linked (Table 2) and R-linked
(Table 4) residues at O-3. First, the acceptors having a D
series peracetylated sugar at O-3 were more reactive toward
glycosylation than the acceptors carrying L series peracety-
lated sugars at O-3. However, as could be expected, the
acceptor carrying an R-linked per-benzylated L-fucose at O-3
was much more reactive than the acetylated analogue. For
the acceptors carrying at O-3 the rather stable β-linked D-
galactosyl and L-fucosyl residues, molecular mechanics cal-
culations suggest that conformational reorientations around
the L-fucosidic linkage and reduction of the conformational
space accessible around this bond contribute to its lower
FIGURE 4. Glycosylation of acceptor 23 and degradation of tri-
saccharide 24: (a) 2 equiv of BF3 Et2O, 40 °C; (b) 2 equiv of
3
BF3 Et2O, 20 °C; (c) 1 equiv of BF3 Et2O, 20 °C; 23 (2), 24 (b),
and 25 (9).
3
3
the diglucosylated trisaccharide 14 synthesized above, and
that was obtained in 27% yield. Thus, the data obtained
from this first reaction suggest that the desired trisaccharide
24 was formed first and very rapidly in these conditions but
then underwent rapid degradation losing the fucosyl residue.
The resulting acceptor disaccharide free at O-3 of the glucos-
amine residue was then glucosylated leading to trisaccharide
14. In an effort to slow down the kinetics of degradation, a
second reaction was carried out at room temperature, using
again 5 equiv of donor 1 and 2 equiv of BF3 Et2O
3
(Figure 4b). Once again the trisaccharide 24 was formed
rapidly and after 50 min of reaction, 70% of the trisaccharide
24 and 30% of residual acceptor 23 were present in the
reaction mixture while no benzylated fucose was observed.
At 100 min of reaction time the ratio of product and acceptor
was only slightly modified in favor of the desired trisacchar-
ide 24 that was formed at 73% while 27% of acceptor 23 was
left in the mixture. Again, no benzylated fucose could be
detected suggesting that degradation of the product and
starting material was slow. However, at 200 min we observed
that the percent abundance of the trisaccharide 24 had
decreased by 22% (to 51%), while the percent abundance
of the acceptor 23 had decreased by 6% (to 21%). At this
time, the hydrolyzed fucosyl peak then accounted for the
remaining 28% of the absorbance at 220 nm. Both reactions
thus show that glucosylation at O-4 of acceptor 23 took
place quite rapidly and much faster than the glucosyla-
tion of acceptor 18 that carried an R-L-peracetylated (vs.
reactivity. Second, in the presence of 2 equiv of BF3 Et2O,
3
the trisaccharide products and/or disaccharide acceptors
that carry R-linked residues at O-3 are prone to degradation
through the loss of this residue while those analogues carry-
ing β-linked peracetylated sugars at O-3 are stable. While
molecular mechanics may suggest an explanation to the
lower reactivity of the β-fucosylated analogue 12 vs. that of
the β-galactosylated acceptor 10, the many stereoelectronic
factors leading to these experimental results cannot be
rationalized rigorously at this level of computation. A more
detailed study involving the modeling of the reaction trajec-
tories and transition states is required to fully understand
these behaviors; however, such computational studies are
(17) Shimizu, H.; Yoshimura, Y.; Hinou, H; Nishimura, S.-I. Tetrahe-
dron 2008, 10091–10096.
8328 J. Org. Chem. Vol. 74, No. 21, 2009

Hendel et al.
JOCArticle
not possible at this time due to the size of the molecules
involved. However, the results reported here clearly illustrate
that the poor yields associated with glycosylation at O-4 of
N-acetylglucosamine acceptors may be due not only to a lack
in the reactivity of the acceptor but also to the instability of
the products in some of the harsher reaction conditions
required to achieve glycosylation. It has been well recognized
that the success of a glycosylation frequently requires that
the donor and acceptor be matched18 based on their respec-
tive expected reactivities19 and configurations (D vs. L and R
vs. β).20 Indeed, it has been well-established that the config-
urations of donors and acceptors as well as the size and
electronic properties of the protecting groups that they carry
had a profound impact on the success of a glycosylation
reaction.18-20 This study now illustrates that the configura-
tion (D vs. L and R vs. β) as well as the protecting group
pattern (Ac vs. Bn) on a sugar unit that is vicinal to the OH
group being glycosylated also has an impact on the success of
the glycosylation reaction. Thus, our results further expand
the match-mismatch concept from the acceptor structural
features to that of the sugar residues that it already carries.
the donor 1 (180 mg, 0.368 mmol, 5.0 equiv) in CH2Cl2 (2.0 mL)
according to the general protocol. Column chromatography
(8:2 EtOAc-hexanes) of the residue gave pure disaccharide 8
(48 mg, 87%) as a colorless glass. [R]D -31 (c 1.0, CH2Cl2); 1H
NMR (400 MHz, CDCl3) δ 7.42-7.28 (m, 5H, Ar), 5.95-5.77
(m, 1H, CH2CHdCH2), 5.75 (d, 1H, J = 8.8 Hz, NH),
5.38-5.18 (m, 2H, CH2CHdCH2), 5.10-4.92 (m, 3H, H-3,
H-30, H-4’), 4.87 (t, 1H, J = 8.3 Hz, H-2’), 4.77-4.59 (m, 2H,
CHHPh, CHHCHdCH2), 4.55-4.41 (m, 4H, H-1, H-10,
CHHPh, CHHCHdCH2), 4.25 (dd, 1H, J = 12.2, 4.7 Hz,
H-6a0), 4.00-3.76 (m, 3H, H-2, H-4, H-6b0), 3.72 (d, 2H, J =
2.4 Hz, H-6a, H-6b), 3.56-3.41 (m, 5H, H-5, H-50, OCH3), 2.04,
1.97, 1.94, 1.90, 1.89 (5s, 15H, 5 ꢀ COCH3); 13C NMR
(100 MHz, CDCl3) δ 170.4, 170.3, 170.2, 169.4, 169.1, 154.8
(CdO), 137.9 (CH2CHdCH2), 131.4, 128.6, 128.5, 128.0, 127.7
(Ar), 118.8 (CH2CHdCH2), 101.5 (C-1), 100.4 (C-10), 76.6
(C-3), 75.8 (C-4), 74.4 (C-5), 73.7 (CH2Ph), 72.9 (C-30), 71.7
(C-50), 71.5 (C-2’), 68.6 (CH2CHdCH2), 68.2 (C-40), 67.5 (C-6),
62.0 (C-60), 56.6 (OCH3), 54.1 (C-2), 23.3, 20.7, 20.6, 20.5
(COCH3); HRESIMS calcd for C34H45NO17 [M þ H]þ
740.2766, found 740.2794.
Methyl 2-Acetamido-4-O-(2,3,4,6-tetra-O-acetyl-β-D-gluco-
pyranosyl)-6-O-benzyl-3-O-trichloroethyloxycarbonyl-2-deoxy-β-
D-glucopyranoside (9). Acceptor 5 (21 mg, 0.043 mmol) was
coupled to the donor 1 (105 mg, 0.215 mmol, 5.0 equiv) in
CH2Cl2 (1.5 mL) according to the general protocol. Column
chromatography (8:2 EtOAc-hexanes) of the residue gave pure
disaccharide 9 (32 mg, 90%) as a colorless glass. [R]D þ9 (c 1.0,
CH2Cl2); 1H NMR (400, CDCl3) δ 7.51-7.31 (m, 5H, Ar), 5.64
(d, 1H, J = 8.5 Hz, NH), 5.14 (dd, 1H, J = 10.1, 8.7 Hz, H-3),
5.09-4.91 (m, 3H, H-30, H-40, CHHCCl3), 4.85 (t, 1H, J = 8.4
Hz, H-20), 4.73 (d, 1H, J = 12.8 Hz, CHHPh), 4.59 (d, 1H, J =
8.0 Hz, H-1), 4.57-4.40 (m, 3H, H-10, CHHPh, CHHCCl3), 4.34
(dd, 1H, J = 12.3, 4.9 Hz, H-6a0), 4.01-3.91 (m, 2H, H-4, H-
6b0), 3.80-3.68 (m, 3H, H-2, H-6a, H-6b), 3.52-3.40 (m, 5H, H-
5, H-50, OCH3), 2.02, 1.94, 1.91, 1.89, 1.87 (5s, 15H, 5 ꢀ
COCH3); 13C NMR (100 MHz, CDCl3) δ 170.6, 170.2, 170.1,
169.4, 169.2, 168.4, 153.8 (CdO), 137.7, 128.6, 128.1, 128.0 (Ar),
101.1 (C-1), 100.2 (C-10), 94.5 (CCl3), 77.3 (C-3), 76.7
(CH2CCl3), 75.6 (C-4), 74.3 (C-5), 73.7 (CH2Ph), 72.9 (C-30),
71.8 (C-50), 71.4 (C-20), 68.1 (C-40), 67.3 (C-6), 62.0 (C-60), 56.8
(OCH3), 54.6 (C-2), 23.3, 20.8, 20.6, 20.5 (COCH3); HRESIMS
calcd for C33H42Cl3NO17 [M þ H]þ 830.1597, found 830.1597.
Methyl 2-Acetamido-3-O-(2,3,4,6-tetra-O-acetyl-β-D-galacto-
pyranosyl)-4-O-(2,3,4,6-tetra-O-acetyl-β-D-glucopyranosyl)-6-
O-benzyl-2-deoxy-β-D-glucopyranoside (13). Alcohol 10 (20 mg,
0.031 mmol) was coupled to the donor 1 (75 mg, 0.152 mmol,
5.0 equiv) in CH2Cl2 (1.0 mL) according to the general protocol.
Column chromatography (9:1 EtOAc:hexanes) of the residue
gave the trisaccharide 13 (28 mg, 93%) as a colorless glass. [R]D
-47 (c 1.0, CH2Cl2); 1H NMR (400 MHz, CDCl3) δ 7.41-7.29
(m, 5H, Ar), 6.46 (d, 1H, J = 6.1 Hz, NH), 5.40 (d, 1H, J =
2.9 Hz, H-40), 5.24-5.12 (m, 2H, H-20, H-300), 5.09-4.99 (m, 2H,
H-30, H-400), 4.94 (dd, 1H, J = 9.8, 8.1 Hz, H-200), 4.81 (d, 1H,
J = 7.9 Hz, H-10), 4.56 (d, 2H, J = 0.9 Hz, CH2Ph), 4.42 (d, 1H,
J = 8.0 Hz, H-100), 4.28 (dd, 2H, J = 12.8, 4.1 Hz, H-1, H-6a00),
4.22-4.07 (m, 4H, H-2, H-5, H-6a0, H-6b0), 4.02-3.93 (m, 3H,
H-3, H-50, H-6b00), 3.86-3.79 (m, 1H, H-4), 3.74 (d, 2H, J = 6.9
Hz, H-6a, H-6b), 3.51-3.44 (m, 1H, H-500), 3.35 (s, 3H, OCH3),
2.11, 2.10, 2.09, 2.08, 2.07, 2.06, 2.05, 2.03, 1.99 (9s, 27H, 9 ꢀ
COCH3); 13C NMR (100 MHz, CDCl3) δ 170.7, 170.5, 170.2,
170.1, 169.9, 169.7, 169.4 (CdO), 138.3, 128.4, 127.9, 127.8 (Ar),
101.4 (C-1), 98.9 (C-10), 98.5 (C-100), 74.5 (C-3), 74.4 (C-4), 73.4
(CH2Ph), 72.6 (C-5), 72.0 (C-500), 71.8 (C-300), 71.3 (C-200), 70.7
(C-30, C-50), 69.8 (C-6), 68.2 (C-20), 68.1 (C-400), 66.9 (C-40), 61.3
(C-600), 60.4 (C-60), 56.1 (OCH3), 48.3 (C-2), 23.4, 23.0, 20.9,
20.8, 20.7, 20.6, 20.5 (COCH3); HRESIMS calcd for
C44H59NO24 [M þ H]þ 986.3505, found 986.3513.
Experimental Section
General Procedure for Glycosylation at O-4 of N-Acetylglu-
cosamine Acceptors. BF3 OEt2 (2.0 equiv) was added to a
3
solution of the acceptor and the known7 glucosyl donor 1 (5.0
equiv) in CH2Cl2 at 40 °C. The reaction mixture was stirred for 1
h, then the reaction was quenched with Et3N (2.3 equiv) and
diluted with CH2Cl2. The mixture was washed with satd aq
NaHCO3, the aqueous layer was re-extracted with CH2Cl2, and
the combined organic layers were dried and concentrated. The
residues were purified by flash chromatography to give the
glycosylated products.
Methyl 2-Acetamido-4-O-(2,3,4,6-tetra-O-acetyl-β-D-gluco-
pyranosyl)-3-O-benzoyl-6-O-benzyl-2-deoxy-β-D-glucopyranoside
(7). Alcohol 3 (15 mg, 0.036 mmol) was coupled tothe donor 1 (92
mg, 0.180 mmol, 5.0 equiv) in CH2Cl2 (1.3 mL) according to the
general protocol. Column chromatography (8:2 EtOAc-
hexanes) of the residue gave pure disaccharide 7 (19 mg, 76%)
1
as a colorless glass. [R]D -13 (c 0.6, CH2Cl2); H NMR (600
MHz, CDCl3) δ 8.01 (d, 2H, J = 7.1 Hz, Ar), 7.60-7.31 (m, 8H,
Ar), 5.59 (d, 1H, J = 9.4 Hz, NH), 5.33 (t, 1H, J = 9.1 Hz, H-3),
4.98 (t, 1H, J = 9.5 Hz, H-3), 4.90-4.70 (m, 3H, H-20, H-40,
CHHPh), 4.58-4.47 (m, 2H, H-10, CHHPh), 4.42 (d, 1H, J = 8.0
Hz, H-1), 4.17 (t, 1H, J = 8.1 Hz, H-2), 4.12 (t, 1H, J = 8.8 Hz,
H-4), 3.89-3.73 (m, 3H, H-6a, H-6b, H-6a0), 3.64 (dd, 1H,
J = 12.2, 2.1 Hz, H-6b0), 3.59-3.45 (m, 4H, H-5, OCH3),
3.31-3.22 (m, 1H, H-50), 1.93, 1.89, 1.84 (3s, 15H, 5 ꢀ COCH3);
13C NMR (150 MHz, CDCl3) δ 170.5, 170.2, 169.3, 169.1, 166.2
(CdO), 137.9, 133.3, 129.8, 129.4, 128.6, 128.4, 128.0, 127.9 (Ar),
102.1 (C-1), 100.0 (C-10), 74.9 (C-4) 74.9 (C-5), 73.8 (CH2Ph),
73.1 (C-3), 72.9 (C-30), 71.7 (C-50), 71.5 (C-20), 67.9 (C-40), 67.6
(C-6), 61.5 (C-60), 56.6 (OCH3), 53.7 (C-2), 23.3, 20.7, 20.6, 20.5
(COCH3);
HRESIMS
calcd
for
C37H45NO16
[M þ H]þ 760.2817, found 760.2791.
Methyl 2-Acetamido-4-O-(2,3,4,6-tetra-O-acetyl-β-D-glucop-
yranosyl)-3-O-allyloxycarbonyl-6-O-benzyl-2-deoxy-β-D-gluco-
pyranoside (8). Acceptor 4 (30 mg, 0.073 mmol) was coupled to
(18) Paulsen, H. In Selectivity a Goal for Synthetic Efficiency; Bartmann,
W., Trost, B. M., Eds; Verlag Chemie: Basel, 1984.
ꢀ
(19) Fraser-Reid, B.; Lopez, J. C.; Gomez, A. M.; Uriel, C. Eur. J. Org.
Chem. 2004, 1387–1395. Uriel, C.; Gomez, A. M.; Lopez, J. C.; Fraser-Reid,
ꢀ
ꢀ
ꢀ
B. Synlett 2003, 14, 2203–2207.
(20) Spijker, N. M.; van Boeckel, C. A. A. Angew. Chem., Int. Ed. Engl.
1991, 30, 180–183.
J. Org. Chem. Vol. 74, No. 21, 2009 8329

Hendel et al.
JOCArticle
Methyl 2-Acetamido-3-O-(2,3,4,6-tetra-O-acetyl-β-D-gluco-
pyranosyl)-4-O-(2,3,4,6-tetra-O-acetyl-β-D-glucopyranosyl)-6-O-
benzyl-2-deoxy-β-D-glucopyranoside (14). Alcohol 11 (16 mg,
0.029 mmol) was coupled to the donor 1 (71 mg, 0.145 mmol,
5.0 equiv) in CH2Cl2 (1.0 mL) according to the general protocol.
Column chromatography (9:1 EtOAc:hexanes) of the residue
gave the trisaccharide 14 (21 mg, 88%) as a colorless glass. [R]D
-62 (c 1.0, CH2Cl2); 1H NMR (400 MHz, CDCl3) δ 7.42-7.29
(m, 5H, Ar), 6.39 (d, 1H, J = 8.0 Hz, NH), 5.26-5.15 (m, 2H,
H-30, H-300), 5.14-5.01 (m, 2H, H-40, H-400), 5.00-4.88 (m, 2H,
H-20, H-200), 4.79 (d, 1H, J = 8.0 Hz, H-100), 4.54 (s, 2H, CH2Ph),
4.43 (d, 1H, J = 8.0 Hz, H-10), 4.38-4.29 (m, 3H, H-1, H-6a0,
H-6a00), 4.13 (dd, 1H, J = 12.5, 2.2 Hz, H-6b00), 4.08-3.92 (m,
4H, H-2, H-3, H-4, H-6b0), 3.82-3.64 (m, 4H, H-5, H-6a, H-6b,
H-50), 3.54-3.48 (m, 1H, H-500), 3.35 (s, 3H, OCH3), 2.05, 2.04,
2.03, 2.02, 1.99, 1.99, 1.98, 1.97 (8s, 27H, 9 ꢀ COCH3); 13C
NMR (100 MHz, CDCl3) δ 171.0, 170.6, 170.4, 169.9 (CdO),
148.1, 138.6, 128.9, 128.4, 128.3 (Ar), 101.6 (C-1), 99.2 (C-10, C-
100), 75.0 (C-5), 73.8 (C-3), 73.2 (CH2Ph), 72.4 (C-4), 72.3, 71.7
(C-30, C-300), 71.6 (C-50, C-500), 70.1 (C-20, C-200), 68.6 (C-6), 68.5
(C-40, C-400), 61.7 (C-60, C-600), 56.5 (OCH3), 50.1 (C-2), 23.5,
21.3, 21.1, 21.0 (COCH3); HRESIMS calcd for C44H59NO24
[M þ Na]þ 1008.3325, found 1008.3318.
H-6a00), 4.13-3.95 (m, 4H, H-3, H-4, H-6b0, H-6b00), 3.82-3.69
(m, 3H, H-2, H-6a, H-6b), 3.53-3.39 (m, 5H, H-5, OCH3, H-
500), 2.16, 2.13, 2.07, 2.03, 2.01, 2.00, 1.99, 1.98, 1.97 (9s, 27H, 9 ꢀ
COCH3); 13C NMR (150 MHz, CDCl3) δ 170.9, 170.6, 170.1,
170.0, 169.7, 169.3 (CdO), 137.9, 128.6, 128.0 (Ar), 102.5 (C-1),
101.1 (C-10), 98.9 (C-100), 77.0 (C-3), 74.7 (C-4), 74.1 (C-5), 73.7
(CH2Ph), 72.7 (C-400), 72.1 (C-500), 71.3 (C-200), 69.5 (C-20), 68.7
(C-30, C-50, C-300), 68.1 (C-6), 66.6 (C-40), 62.8 (C-600), 62.1
(C-60), 56.7 (OCH3), 54.1 (C-2), 23.5, 21.0, 20.8, 20.7,
20.6 (COCH3); HRESIMS calcd for C44H59NO24 [M þ H]þ
986.3505, found 986.3533.
Methyl 2-Acetamido-3-O-(2,3,4-tri-O-acetyl-r-L-fucopyano-
syl)-4-O-(2,3,4,6-tetra-O-acetyl-β-D-glucopyranosyl)- 6-O-ben-
zyl-2-deoxy-β-D-glucopyranoside (21). Alcohol 18 (13 mg,
0.021 mmol) was coupled to the donor 1 (60 mg, 0.123 mmol,
5.0 equiv) in CH2Cl2 (0.8 mL) according to the general protocol.
Column chromatography (10:1 EtOAc:hexanes) of the residue
gave an inseparable mixture of the acceptor 18 and trisaccharide
21 (10 mg). The yields of 18 and 21 in the mixture were
determined by NMR to be 34% and 26%, respectively. An
analytical sample of 21 was obtained after additional column
chromatography (8:2 CH2Cl2:acetone). [R]D -59 (c 0.4, CH2-
Cl2); 1H NMR (400 MHz, CDCl3) δ 7.43-7.31 (m, 5H, Ar), 5.56
(d, 1H, J = 8.5 Hz, NH), 5.34 (d, 1H, J = 3.8 Hz, H-10), 5.28 (dd,
1H, J = 10.8, 3.2 Hz, H-20), 5.19 (d, 1H, J = 2.8 Hz, H-40),
5.03-4.87 (m, 3H, H-30, H-400, H-300), 4.78 (t, 1H J = 8.6 Hz,
H-200), 4.75-4.63 (m, 2H, CHHPh, H-50), 4.61-4.38 (m, 4H,
H-1, H-6a00, CHHPh, H-100), 4.06-3.89 (m, 2H, H-3, H-4,
H-6b00), 3.84-3.69 (m, 2H, H-6a, H-6b), 3.64 (q, 1H, J =
7.0 Hz, H-2), 3.49-3.39 (m, 1H, H-5), 3.36 (s, 3H, OCH3),
3.31-3.21 (m, 1H, H-500), 2.06, 2.01, 2.00, 1.99, 1.94, 1.93, 1.90,
1.89 (8s, 24H, 8 ꢀ COCH3), 1.07 (d, 3H, J = 6.5 Hz, H-60); 13C
NMR (150 MHz, CDCl3) δ 171.1, 170.7, 170.5, 170.2, 169.5,
169.4, 169.1(CdO), 137.8, 128.7, 128.5, 128.4, 128.1, 128.0,
127.8 (Ar), 100.9 (C-1), 99.6 (C-100), 95.1 (C-10), 74.6 (C-3),
74.5 (C-5), 73.6 (CH2Ph), 73.3 (C-500), 72.7 (C-4), 72.2 (C-30),
71.7 (C-40), 71.1 (C-200), 68.6 (C-400), 68.0 (C-6), 67.7 (C-300), 67.6
(C-20), 64.5 (C-50), 61.0 (C-600), 56.6 (OCH3), 56.3 (C-2), 23.4,
21.0, 20.7, 20.6 (COCH3), 15.7 (C-60); HRESIMS calcd for
C42H57NO22 [M þ Na]þ 950.3270, found 950.3266.
Methyl 2-Acetamido-3-O-(2,3,4-tri-O-acetyl-β-L-fucopyrano-
syl)-4-O-(2,3,4,6-tetra-O-acetyl-β-D-glucopyranosyl)-6-O-benzyl-
2-deoxy-β-D-glucopyranoside (15). Alcohol 12 (13 mg, 0.021
mmol) was coupled to the donor 1 (52 mg, 0.106 mmol,
5.0 equiv) in CH2Cl2 (0.8 mL) according to the general protocol.
Column chromatography (9:1 EtOAc:hexanes) of the residue
gave an inseparable mixture of the acceptor 12 and trisaccharide
15 (16 mg). The yields of 12 and 15 in the mixture were
determined by NMR to be 24% and 63%, respectively. An
analytical sample of trisaccharide 15 was obtained upon addi-
tional column chromatography (25:1 CHCl3:MeOH). [R]D -5
(c 0.4, CH2Cl2); 1H NMR (400 MHz, CDCl3) δ 7.44-7.30 (m,
5H, Ar), 5.98 (d, 1H, J = 8.1 Hz, NH), 5.29 (d, 1H, J = 3.4 Hz,
H-40), 5.19-4.99 (m, 4H, H-20, H-30, H-300, H-400), 4.93 (d, 1H,
J = 8.2 Hz, H-200), 4.82 (d, 1H, J = 7.8 Hz, H-10), 4.71-4.50 (m,
4H, H-1, H-100, CH2Ph), 4.24 (dd, 1H, J = 12.4, 4.0 Hz, H-6a00),
4.14 (t, 1H, J = 6.8 Hz, H-3), 4.03-3.94 (m, 2H, H-4, H-6b00),
3.86 (q, 1H, J = 6.8 Hz, H-50), 3.83-3.70 (m, 3H, H-2, H-6a,
H-6b), 3.64-3.55 (m, 1H, H-5), 3.48 (s, 3H, OCH3), 3.44-3.38
(m, 1H, H-500), 2.18, 2.08, 2.04, 2.03, 1.98 (5s, 24H, 8 ꢀ COCH3),
1.24 (d, 3H, J = 6.4 Hz, H-60); 13C NMR (100 MHz, CDCl3) δ
171.0, 170.5, 170.4 (CdO), 129.0, 128.4, 128.3 (Ar), 102.1 (C-1),
100.6 (C-10), 99.0 (C-100), 76.8 (C-3), 75.3 (C-5), 75.0 (C-4), 73.9
(CH2Ph), 72.9 (C-300), 72.4 (C-500, 72.0 (C-200), 71.6 (C-30), 70.8
(C-40), 70.0 (C-20), 69.8 (C-50), 69.4 (C-6), 68.5 (C-400), 61.8
(C-600), 57.0 (OCH3), 54.2 (C-2), 23.8, 21.4, 21.2, 21.1,
21.0 (COCH3), 16.4 (C-60); HRESIMS calcd for C42H57NO22
[M þ H]þ 928.3424, found 928.3434.
Methyl 2-Acetamido-4-O-(2,3,4,6-tetra-O-acetyl-β-D-glucop-
yranosyl)-3-O-(2,3,4-tri-O-acetyl-r-L-rhamnopyranosyl)-6-O-
benzyl-2-deoxy-β-D-glucopyranoside (22). Alcohol 19 (14 mg,
0.029 mmol) was coupled to the donor 1 (56 mg, 0.114 mmol,
5.0 equiv) in CH2Cl2 (0.8 mL) according to the general protocol.
Column chromatography (9:1 EtOAc:hexanes) of the residue
gave an inseparable mixture of the acceptor 19 and trisaccharide
22 (12 mg). The yields of 19 and 22 in the mixture were
determined by NMR to be 43% and 27%, respectively. An
analytical sample of 22 was obtained after additional column
chromatography (8:2 CH2Cl2:acetone). [R]D -9 (c 0.8, CH2Cl2);
1H NMR (400 MHz, CDCl3) δ 7.51-7.31 (m, 5H, Ar), 6.07 (d,
1H, J = 8.6 Hz, NH), 5.27-5.19 (m, 2H, H-20, H-30), 5.16-5.01
(m, 2H, H-40, H-300), 4.98 (s, 1H, H-10), 4.95-4.77 (m, 4H, H-1,
CHHPh, H-200, H-400), 4.61 (d, 1H, J = 8.2 Hz, H-100), 4.53 (d,
1H, J = 11.9 Hz, CHHPh), 4.46-4.36 (m, 1H, H-50), 4.32-4.10
(m, 3H, H-3, H-6a00, H-6b00), 3.96 (t, 1H, J = 7.6 Hz, H-4), 3.82
(d, 2H, J = 3.3 Hz, H-6a, H-6b), 3.61-3.54 (m, 1H, H-5),
3.53-3.40 (m, 5H, H-2, H-500, OCH3), 2.10, 2.08, 2.05, 2.00,
1.97, 1.96, 1.95, 1.94 (8s, 24H, 8 ꢀ COCH3), 1.24 (d, 3H, J =
6.3 Hz, H-60); 13C NMR (100 MHz, CDCl3) δ 172.8, 171.1,
170.7, 170.4, 170.2, 169.9, 169.5, 169.3 (CdO), 128.6, 128.1,
128.0 (Ar), 100.2 (C-1), 98.9 (C-100), 96.3 (C-10), 74.2 (C-5), 73.8
(C-4), 73.6 (CH2Ph), 73.1 (C-3), 72.7 (C-300), 71.8 (C-500), 71.4
(C-400), 71.0 (C-40), 70.1 (C-30), 69.3 (C-200), 69.0 (C-20), 68.2
(C-6), 66.3 (C-50), 62.7 (C-600), 56.8 (C-2), 56.8 (OCH3), 23.4,
21.0, 20.8, 20.7, 20.6 (COCH3), 17.1 (C-60); HRESIMS calcd for
C42H57NO22 [M þ H]þ 928.3450, found 928.3492.
Methyl 2-Acetamido-4-O-(2,3,4,6-tetra-O-acetyl-β-D-glucop-
yranosyl)-3-O-(2,3,4,6-tetra-O-acetyl-r-D-mannopyranosyl)-6-
O-benzyl-2-deoxy-β-D-glucopyranoside (20). Alcohol 17 (22 mg,
0.034 mmol) was coupled to the donor 1 (83 mg, 0.167 mmol,
5.0 equiv) in CH2Cl2 (1.2 mL) according to the general protocol.
Column chromatography (10:1 EtOAc:hexanes) of the residue
gave a colorless glass (21 mg) containing the trisaccharide 20
contaminated with some degraded donor. The yield of trisac-
charide 20 (18 mg, 55%) was calculated from the 1H NMR
spectrum while an analytical sample of trisaccharide 20 was
obtained after additional column chromatography (35:1 CHCl3:
MeOH). [R]D þ5 (c 0.6, CH2Cl2); 1H NMR (600 MHz, CDCl3)
δ 7.48-7.30 (m, 5H, Ar), 5.98 (d, 1H, J = 8.7 Hz, NH), 5.43
(s, 1H, H-20), 5.33-5.21 (m, 2H, H-30, H-40), 5.12 (d, 1H, J =
1.4 Hz, H-10), 5.09-4.96 (m, 2H, H-300, H-400), 4.89 (t, 1H, J =
8.1 Hz, H-200), 4.75 (d, 1H, J = 11.9 Hz, CHHPh), 4.60-4.47
(m, 3H, H-1, CHHPh, H-100), 4.34-4.13 (m, 3H, H-50, H-6a0,
8330 J. Org. Chem. Vol. 74, No. 21, 2009

Hendel et al.
JOCArticle
Experimental Methods for Analytical and Preparative HPLC
during the Glucosylation of Disaccharide Acceptor 23. An ali-
quot (5 μL) of the reaction mixture was diluted in CH3CN
(100 μL) containing NEt3 and 10 μL of this solution was
injected in the HPLC. Analytical HPLC analysis was carried
out on a Novapak C18 reverse phase column (3.9 ꢀ 150 mm)
eluted with a gradient solvent system of CH3CN:H2O (65:35
to 90:10) at a 2 mL/min flow rate. Detection was carried out
with a UV detector set at 220 nm. The trisaccharide 24,
unreacted acceptor 23, and fucose 28 were separated
by RP-HPLC on a C18 column (8 ꢀ 200 mm) eluted (2 mL/
min) with the gradient solvent system mentioned above over
15 min.
Acknowledgment. We thank the National Science and
Engineering Research Council of Canada, the Canada
Foundation for Innovation, and the Ontario Innovation
Trust for financial support of this work.
Supporting Information Available: The Results and Discus-
sion and Experimental Section for the Synthesis of the Glycosyl
Acceptors 4, 5, 11, 12, 17, and 18, as well as the General
Experimental Methods and the Computational Methods are
1
given, along with H, COSY, and 13C NMR data for all new
compounds 4, 5, 7-9, 11-15, 17, 18, 20-22, 31, 32, 35, 36, 38,
40. This material is available free of charge via the Internet at
http://pubs.acs.org.
J. Org. Chem. Vol. 74, No. 21, 2009 8331