Study of interhalogens/silver trifluoromethanesulfonate as promoter systems for high-yielding sialylations

Article abstract of DOI:10.1021/jo0262412

We have studied interhalogen/silver trifluoromethanesulfonate (IX/AgOTf) promoted glycosylations and found differences in the sensitivity of the formed oxocarbenium ions (e.g. from compounds with or without participating groups) toward halide nucleophiles. These differences can be explained using the HSAB theory. By applying this theory on sialylations, we increased the yield for a model reaction from a highly unpredictable 35-46% using ICl to 74% using IBr. We have also showed that the most prominent role of the silver ions is lowering the concentration of the halide nucleophile rather than activating the interhalogen compound and, by increasing the amount of AgOTf from 1 to 1.5 equiv (with respect to IBr), the yield in the model reaction improved from 74% to 89%. A comparison of two different anomeric leaving groups showed that glycal formation can be minimized using a thiophenyl donor instead of xanthate. By combining these observations, we were able to increase the yield of the model reaction to 97%.

Full text of DOI:10.1021/jo0262412

Stu d y of In ter h a logen s/Silver Tr iflu or om eth a n esu lfon a te a s
P r om oter System s for High -Yield in g Sia lyla tion s
Andre´as Meijer and Ulf Ellervik*
Organic and Bioorganic Chemistry, Center for Chemistry and Chemical Engineering, Lund University,
P.O. Box 124, SE-221 00 Lund, Sweden
ulf.ellervik@bioorganic.lth.se
Received J uly 24, 2002
We have studied interhalogen/silver trifluoromethanesulfonate (IX/AgOTf) promoted glycosylations
and found differences in the sensitivity of the formed oxocarbenium ions (e.g. from compounds
with or without participating groups) toward halide nucleophiles. These differences can be explained
using the HSAB theory. By applying this theory on sialylations, we increased the yield for a model
reaction from a highly unpredictable 35-46% using ICl to 74% using IBr. We have also showed
that the most prominent role of the silver ions is lowering the concentration of the halide nucleophile
rather than activating the interhalogen compound and, by increasing the amount of AgOTf from
1 to 1.5 equiv (with respect to IBr), the yield in the model reaction improved from 74% to 89%. A
comparison of two different anomeric leaving groups showed that glycal formation can be minimized
using a thiophenyl donor instead of xanthate. By combining these observations, we were able to
increase the yield of the model reaction to 97%.
SCHEME 1. Activa tion a n d Rea ction P a th w a y for
Sia lic Acid Th io Don or s Usin g ICl/AgOTf
In tr od u ction
The importance of sialic acid containing oligosaccha-
rides and glycoconjugates is reflected in the immense
number of different methods for sialylation found in the
recent literature.1 However, despite numerous new meth-
ods, no approach has been reported that allows glycosyl-
ation of a wide range of acceptors in high yields and
stereoselectivities.1 In a recent paper we presented iodine
monochloride/silver trifluoromethanesulfonate (ICl/Ag-
OTf) as a convenient and efficient promoter system for
thioglycoside activation,2 and in a number of glycosyla-
tions we demonstrated its capability. ICl/AgOTf, like
most promoter systems for thioglycoside activation (e.g.
MeSBr/AgOTf or NIS/TfOH),3 first generates a thiophilic
species (i.e. an iodonium ion, “I+”) that activates the
sulfur-containing anomeric leaving group to generate an
electrophilic oxocarbenium ion (Scheme 1). The ICl/
AgOTf system worked well for most glycosylations but
gave low yields and unpredictable results upon sialyla-
tion (Scheme 2 and Table 1; entry 1). This paper describes
mechanistic details and optimization of interhalogen/
silver trifluoromethanesulfonate promoted sialylations.
glycoside 6, sialic acid glycal 4,5 and, more surprisingly,
sialic acid chloride 35 (Scheme 1).
To examine if sialic acid chloride 3 is a stable byproduct
or a slowly reacting intermediate, we synthesized the
chloride 3 and subjected it to AgOTf under standard
sialylation conditions (Table 1; entries 7 and 8). No
disaccharide 6 could be isolated from the reaction mix-
ture; only acceptor 5 and sialic acid chloride 3 together
with decomposed donor in the form of sialic acid glycal 4
were recovered.
Resu lts a n d Discu ssion
To investigate why sialic acid donor 14 performed
poorly under ICl/AgOTf promoted conditions,2 we con-
ducted an NMR experiment which indicated that the
donor is activated and consumed within 1 h, forming
(1) Boons, G. J .; Demchenko, A. V. Chem. Rev. 2000, 100, 4539.
(2) Ercegovic, T.; Meijer, A.; Magnusson G.; Ellervik, U. Org. Lett.
2001, 3, 913.
(3) (a) Veeneman, G. H.; Van Leeuwen, S. H.; Van Boom, J . H.
Tetrahedron Lett. 1990, 31, 1331. (b) Dasgupta, F.; Garegg, P. J .
Carbohydr. Res. 1988, 177, C13.
The formation of the chloride 3, despite an equimolar
amount of silver ions added,6 indicates that sialylation
(5) Byramova, N. E.; Tuzikov, A. B.; Bovin, N. V. Carbohydr. Res.
1992, 237, 161 and references therein.
(4) Marra, A.; Sinay, P. Carbohydr. Res. 1989, 187, 35.
10.1021/jo0262412 CCC: $22.00 © 2002 American Chemical Society
Published on Web 09/24/2002
J . Org. Chem. 2002, 67, 7407-7412
7407

Meijer and Ellervik
SCHEME 2 a
a
Reaction conditions are given in Table 1.
TABLE 1. Rea ction Da ta for Sch em es 2 a n d 3
entry
donor
acceptor
IX
mole ratioa
reaction conditionsb
productc
yield (%)
1
2
3
4
5
6
7
8
14
1
1
1
1
512
5
5
5
5
5
5
5
5
ICl
IBr
I2
IBr
ICl
IBr
ICl
1.5:1.0:1.5:1.5
1.5:1.0:1.5:1.5
1.5:1.0:1.5:1.5
1.5:1.0:2.0:3.0
1.5:1.0:1.5:-
1.5:1.0:1.5:-
1.5:1.0:1.5:1.5
1.5:1.0:-:1.5
1.5:1.0:-:1.5
1.3:1.0:-:3.0
1.3:1.0:1.3:1.3
1.3:1.0:1.3:1.3
1.5:1.0:1.5:1.5
1.5:1.0:1.5:1.5
1.5:1.0:2.0:3.0
3.0:1.0:3.2:6.0
1.5:1.0:2.0:3.0
3.0:1.0:3.2:6.0
1.5:1.0:2.0:3.0
3:2 MeCN-CH2Cl2, -72 °C, 2 h
3:2 MeCN-CH2Cl2, -72 °C, 2 h
3:2 MeCN-CH2Cl2, -72 °C, 2 h
3:2 MeCN-CH2Cl2, -72 °C, 2 h
3:2 MeCN-CH2Cl2, -72 °C, 2 h
3:2 MeCN-CH2Cl2, -72 °C, 2 h
3:2 MeCN-CH2Cl2, -72 °C, 2 h
3:2 MeCN-CH2Cl2, -72 °C, 2 h
3:2 MeCN-CH2Cl2, -72 °C, 2 h
1:2 MeCN-CH2Cl2, -72°C, 1 h
3:2 MeCN-CH2Cl2, -72 °C, 2 h
3:2 MeCN-CH2Cl2, -72 °C, 2 h
3:2 MeCN-CH2Cl2, -72 °C, 2 h
3:2 MeCN-CH2Cl2, -72 °C, 2 h
3:2 MeCN-CH2Cl2, -72 °C, 2 h
3:2 MeCN-CH2Cl2, -72 °C, 2 h
3:2 MeCN-CH2Cl2, -72 °C, 2 h
3:2 MeCN-CH2Cl2, -72 °C, 2 h
3:2 MeCN-CH2Cl2, -72 °C, 2 h
6
6
6
6
6
6
35-46
74
67
89
0d
1
0d
35
3
0
0
0
9
75
813
1015
10
211
2
10
11
12
13
14
15
16
17
18
19
914
1115
11
5
0
ICl
IBr
ICl
IBr
IBr
IBr
IBr
IBr
IBr
1215
12
6
6
6
14
1632
14
16
0d
35
50
87
97
43
75
44
88
5
5
2
1
1
2
1325
1531
13
15
2
a
b
Donor:acceptor:IX:AgOTf. Solvents and molecular sieves were dried and activated, respectively, using conventional methods. c NMR
d
data were in agreement with those reported in the literature. Only trace amounts of disaccharidic product isolated.
using interhalogens is a more complex reaction than we
originally anticipated. We suggest that a polyhalide ion,
ICl2-, which can be formed from Cl- reacting with ICl,7,8
is the actual chloride nucleophile. The ICl2-, generated
through the dissociation of ICl in MeCN,9 does not form
a precipitate of AgCl when subjected to AgOTf. Instead,
-
UV spectroscopy showed a lowered concentration of ICl2
,
(6) Luehrs, D. C.; Iwamoto, R. T.; Kleinberg, J . J . Inorg. Chem. 1966,
5, 201.
(7) (a) Popov, A. I.; Pflaum, R. T. J . Am. Chem. Soc. 1957, 79, 570.
(b) Faull, H. J . J . Am. Chem. Soc. 1934, 56, 522.
(8) Popov, A. I.; Swensen, R. F. J . Am. Chem. Soc. 1955, 77, 3724.
(9) Greenwood, N. N.; Earnshaw, A. Chemistry of the Elements;
Pergamon Press: Oxford, U.K., 1997; p 824.
7408 J . Org. Chem., Vol. 67, No. 21, 2002

Interhalogens/ Silver Trifluoromethanesulfonate
F IGUR E 1. Plot showing the absorption spectra of an ICl
solution titrated with AgOTf. The absorption maximum at 227
- 8
nm is the characteristic peak of ICl2
.
indicating the formation of a soluble complex between
-
silver ions and ICl2 (Figure 1).
To further demonstrate the nucleophilicity of an ICl
solution, we performed a simple NMR experiment where
methyl iodide (MeI) and ICl were dissolved in deuterated
acetonitrile. Without silver ions present, the formation
of methyl chloride is reasonably fast, having second-order
reaction kinetics with respect to MeI and ICl (Figure 2a).
When 0.5 equiv of AgOTf is added, the reaction slows
down and proceeds with more complex reaction kinetics
(Figure 2b).
Once we isolated the problem to interfering chloride
nucleophiles, we turned our attention to the more in-
triguing question of why some glycosylations work well
whereas sialylations may be inferior, when using ICl/
AgOTf.2
F IGURE 2. (a) Plot showing second-order reaction kinetics
for the nucleophilic substitution of MeI using ICl. (b) Plot
showing slower rate and more complex reaction kinetics when
0.5 equiv of AgOTf is added.
reaction outcome on sugars other than sialic acid, we
subjected the sulfide donor 1015 to the ICl/AgOTf pro-
moter system. This reaction gave no disaccharidic prod-
uct but, instead, the corresponding chloride16 in addition
to recovered acceptor 1115 (Table 1; entry 11).
In summary, the absence of a participating group
makes the activated donor more reactive toward the
competing reaction with nucleophilic chloride species,
forming a stable sugar halide and thereby lowering the
total yield of the reaction.
The explanation for the differences in chloride affinities
might be that the oxocarbenium ion formed on activation
of the donor without a participating group is a hard
electrophile (Figure 3a), well-matched to the chloride
nucleophile.17 In comparison, the softer acyloxonium ion
formed by a participating group (Figure 3b) will interact
less with the hard chloride nucleophile. The possibility
for some delocalization by interaction with the ester
functionality (Figure 3c)18 renders the sialic acid less
hard than a standard glycoside without a participating
group.19
Since the donors that performed well (i.e. did not form
glycosyl chlorides) with the ICl/AgOTf system all carried
a participating group and the sialic acid donor 1 did not,2
we now hypothesized that the interaction of the partici-
pating groups with the anomeric position of the donor
influenced the reaction outcome. The interaction might
result in one of the following scenarios: (i) the glycosyl
chloride formed from a donor carrying a participating
group is activated under these reaction conditions or (ii)
glycosyl donors carrying a participating group are less
prone to be transformed into chlorides.
To examine scenario i, i.e., if glycosyl chlorides are
reactive intermediates, chloride donor 813 was mixed with
AgOTf and acceptor 914 (Table 1; entry 10). The thio
equivalent of donor 8 is capable of glycosylating acceptor
9 using ICl/AgOTf at the same temperature.2 However,
no product was obtained, implying that this donor
carrying a participating group is not activated under
these reaction conditions (-72 °C).13 Then, to examine
whether the absence of participating groups influence the
(10) Boons, G. J . Carbohydrate Chemistry; Blackie Academic: Lon-
don, 1998; p 139.
(11) Kirchner, E.; Thiem, F.; Dernick, R.; Heukeshoven, J .; Thiem,
J . J . Carbohydr. Chem. 1988, 7, 453.
The different selectivities of donors with or without
participating groups have been noted by others. Fraser-
(12) Wilstermann, M.; Kononov, L. O.; Nilsson, U.; Ray, A. K.;
Magnusson, G. J . Am. Chem. Soc. 1995, 117, 4742.
(13) Donor 8 has been used in a glycosylation reaction with AgOTf
as promoter, albeit at a higher temperature (-45 °C to room temper-
ature): Pozsgay, V.; Trinh, L.; Shiloach, J .; Robbins, J . B.; Donohue-
Rolfe, A.; Calderwood, S. B. Bioconjugate Chem. 1996, 7, 45.
(14) Ellervik, U.; Magnusson, G. Carbohydr. Res. 1996, 280, 251.
(15) Wilstermann, M.; Balogh, J .; Magnusson, G. J . Org. Chem.
1997, 62, 3659.
(16) Kova´c, P. J . Carbohydr. Chem. 1992, 11, 999.
(17) Hevesi, L. Phosphorus, Sulfur Silicon Relat. Elem. 2001, 171,
55.
(18) Horenstein, B. A. J . Am. Chem. Soc. 1997, 119, 1101.
(19) Ho, T. Chem. Rev. 1975, 75, 1.
J . Org. Chem, Vol. 67, No. 21, 2002 7409

Meijer and Ellervik
was synthesized and mixed with AgOTf and acceptor 5
at -72 °C (Table 1; entry 9). No product 6 was formed;
i.e., the increased yield is due to less interaction between
the bromide ion and the carbocation.
To investigate the role of silver ions, we found ICl so
reactive that it can partly activate the sialic acid donor
1 under our sialylation conditions even without silver
present (Table 1; entry 5).24 However, only traces of
disaccharidic product were detected, with sialic acid
glycal 4 and chloride 3 as major products, indicating the
importance of silver ions in lowering the concentration
of competing chloride nucleophiles. IBr is also capable
of activating xanthate donor 1 under the same conditions,
but less donor is consumed after the same reaction time
(Table 1; entry 6). Molecular iodine is too unreactive and
not capable of activating all donor 1 even with silver
present (Table 1; entry 3).
Logically, an increase of the silver ion concentration
would lower the amount of the halide nucleophile and
thereby increase the yield of the sialylation. Indeed, when
the amount of AgOTf is increased from 1 to 1.5 equiv
(with respect to IBr), the yield in the model reaction
improved from 74% to 89% (Table 1; entry 4) without
even trace amounts of sialic acid bromide 7.25
F IGURE 3. (a) Possible stabilization of an activated glycosyl
donor by resonance. (b) Possible stabilization of an activated
glycosyl donor by resonance and formation of an acyloxonium
ion. (c) Possible stabilization of an activated sialic acid donor
by resonance and interactions with the ester functionality.
In addition to formation of sialic acid halides, sialic acid
glycal 4 is also formed as a byproduct in the reaction. To
evaluate the importance of the leaving group on the
formation of different byproducts, and to confirm its
redundancy in the interaction between interhalogen
nucleophiles and activated sialic acid species, we sub-
jected sialic acid thiophenyl donor 211 to IX/AgOTf
conditions (Table 1; entries 13 and 14). The xanthate
donor 1 is usually considered as a more powerful donor
in sialylations,26 but to our surprise we obtained better
yields using the thiophenyl donor: 50% vs 35-46% using
ICl and 87% vs 74% using IBr.
Reid reported a glycosylation reaction where the pres-
cence or absence of a participating group on the donor
controlled the regioselectivity. The selectivity could not
be explained by purely steric factors. Instead, a model
describing the electronic properties of the carbocations
(depicted as Diffuse and Compact) was used to explain
the reaction outcome.20
One way to minimize the unwanted interactions would
be to change the promoter system to IBr or I2/AgOTf, thus
changing the nucleophilic species from chloride to the
softer bromide or iodide, which would interact less with
the electrophile.
Indeed, glycosylation of acceptor 11 with the hard
donor 10, using the softer promoter system IBr/AgOTf,
this time yielded disaccharide 12 (Table 1; entry 12).21
Sialylation of acceptor 5 with donor 1 using IBr/AgOTf
produced 6 in the improved yield of 74% (Table 1; entry
2), whereas I2/AgOTf gave a somewhat lower yield of 67%
(Table 1; entry 3) together with unactivated donor,
indicating the low reactivity of molecular iodine. These
results are in full agreement with the reactivity order of
the interhalogens that has been thoroughly investigated
by Field and established as ICl > IBr > I2.22 To prove
that the increased yield in the IBr promoted reaction is
not due to the transformation of sialic acid bromide 75 to
disaccharide 6 by the action of silver ions,23 bromide 7
To learn more about the reactions, we perfomed a study
of the reaction kinetics. Small-scale reactions were
conducted and stopped at different times (using cyclo-
hexene and diisopropylamine) and then analyzed by
NMR. Interestingly, the xanthate donor seems to follow
a faster initial reaction rate than the thiophenyl donor.
After 1 min 50% of the xanthate donor is consumed,
compared to 15% of the thiophenyl donor (Figure 4).
However, while the thiophenyl donor is consumed in
about 10 min, the activation of the xanthate donor slows
down and trace amounts of unreacted donor can be
isolated even after 1 h.27 We therefore speculate that the
higher yield using the thiophenyl donor compared to the
xanthate donor is due to the different rates of consump-
tion (i.e. formation of reactive species). The fast initial
activation of the xanthate donor gives rise to a higher
(24) Field has successfully used the interhalogens as promoters
without the aid of AgOTf. However these glycosylation reactions were
carried out at higher temperatures (-10 °C): (a) Kartha, K. P. R.; Field,
R. A. Tetrahedron Lett. 1997, 38, 8233. (b) Kartha, K. P. R.; Aloui, M.;
Field, R. A. Tetrahedron Lett. 1996, 37, 5175.
(25) We have been unable to isolate sialic acid bromide 7 due to its
instability and subsequent decomposition to glycal 4. However, we have
been able to isolate the ethyl sialoside formed from the reaction
between sialic acid bromide 7 and EtOH from the purification system
at workup.33
(20) (a) Anilkumar, G.; Nair, L. G.; Fraser-Reid, B. Org. Lett. 2000,
17, 2587. (b) Fraser-Reid, B.; Lopez, J . C.; Radhakrishnan, K. V.; Mach,
M.; Schlueter, U.; Gomez, A. M.; Uriel, C. J . Am. Chem. Soc. 2002,
124, 3198.
(21) The yield has increased from 0% (ICl) to 35% (IBr) but is still
modest. We wish to point out that the reaction conditions are not
optimized but chosen for best comparison with other glycosylations
reported in this paper.
(22) Cura, P.; Aloui, M.; Kartha, K. P. R.; Field, R. A. Synlett 2000,
9, 1279.
(23) Paulsen, H.; Von Deessen, U. Carbohydr. Res. 1988, 175, 283.
(26) (a) Birberg, W.; Lo¨nn, H. Tetrahedron Lett. 1991, 32, 7453. (b)
Marra, A.; Sinay¨, P. Carbohydr. Res. 1990, 195, 303.
7410 J . Org. Chem., Vol. 67, No. 21, 2002

Interhalogens/ Silver Trifluoromethanesulfonate
SCHEME 3 a
F IGURE 4. Plot showing the amount of unreacted donor left
in the reaction mixture as a function of time: (-9-) xanthate
donor 1; (-b-) thiophenyl donor 2.
concentration of oxocarbenium ions, which we believe
explains the higher proportion of sialic acid glycal using
this donor.
As a final experiment, we subjected the thiophenyl
donor to the higher silver ion concentrations and, indeed,
increased the yield of the model reaction from 89% to 97%
(Table 1; entry 15).
Two more acceptors were sialylated in order to prove
the usefulness of the new promoter system. The chal-
lenging acceptor 1331 was sialylated using 3 equiv of
donor 1 or 2 in excellent yields of 43% and 44%,
respectively, with an R: â:lactone ratio of 66:17:17 (Scheme
3 and Table 1; entries 16 and 18). Prior attempts to
sialylate acceptor 13 with donor 1 using MeSBr/AgOTf
gave no product.31 The lactose acceptor 1532 was sialyl-
ated in 75% and 88% yields (R: â 95:5) using 1.5 equiv of
xanthate donor 1 or thiophenyl donor 2 (Table 1; entries
17 and 19).30
a
Reaction conditions are given in Table 1.
nucleophilic halide species in the glycosylation reaction
follow the predictions of the HSAB theory,19 which
isessentially a new concept in synthetic carbohydrate
chemistry.
The halide formation can be minimized by using IBr
instead of ICl, and we present IBr/AgOTf as a convenient
and efficient promoter system for high-yielding sialyla-
tion reactions. Our investigations, which include some
mechanistic details of the reaction, have also shown that
the most prominent role of the silver ions is lowering the
concentration of halide nucleophile rather than activating
the interhalogen compound. This opens the door to future
investigations using other halophiles.
Con clu sion
To summarize, we have investigated interhalogen/
silver trifluoromethanesulfonate (IX/AgOTf) promoted
sialylation reactions. We conclude that formation of sialic
acid halide and sialic acid glycal are the two major causes
for low yields in the reaction.
We have shown that the interactions of oxocarbenium
ions, formed from different carbohydrate donors, with
Formation of sialic acid glycal can be minimized by
using a thiophenyl donor with a slower initial reaction
rate than the corresponding xanthate donor. By a com-
bination of these findings, we were able to increase the
yield of the model reaction from 46% to 97%.
The stereoselectivities for the ICl-, IBr-, and I2-
promoted sialylations of acceptor 5 (R: â 8:1) were virtu-
ally identical (Table 1; entries 1-4 and 13-15), and the
R/â selectivity for the sialylation of lactose acceptor 15
using IBr as promoter (Table 1; entries 17 and 19) is
comparable to published results, using PhSCl/AgOTf as
promoter.30 This suggests that the choice of leaving group
and thiophilic species in the promoter system is of minor
importance for the stereochemical outcome of the glyco-
sylation.34
(27) We have been able to isolate dixantogen28 (the disulfide of ethyl
xanthate) and phenyl disulfide29 in the respective donor reaction
mixtures. The isolation of disulfides indicates that an electrophilic
sulfur species is present in the reaction mixture.30 We postulate that
the species responsible is the sulfenyl iodide formed from the reaction
between iodonium ion (from IX/AgOTf) and thio donor.3a Thus, the role
of the sulfur leaving group is not only to be reactive enough to be
activated by the thiophilic species but also to form a “second” thiophilic
species in the reaction mixture. These findings might explain the
different consumption patterns of the xanthate and thiophenyl donor
seen in Figure 4. Further investigations regarding the kinetics of
iodonium-promoted activation of thioglycosides are in progress.
(28) (a) Cox, B. G.; Natarajan, R. J . Chem. Soc., Perkin Trans. 2
1977, 113. (b) Barany, G.; Schroll, A. L., Mott, A. W.; Halsrud, D. A.
J . Org. Chem. 1983, 48, 4750.
(29) Pouchert, C. J .; Behnke, J . The Aldrich Library of 13C and 1H
FT NMR Spectra, 1st ed.; Aldrich: Milwaukee, WI, 1993; Vol. 2, 440c.
(30) Martichonok, V.; Whitesides, G. M. J . Org. Chem. 1996, 61,
1702.
(31) Ellervik, U.; Grundberg, H.; Magnusson, G. J . Org. Chem. 1998,
63, 9323.
(32) J ansson, K.; Ahlfors, S.; Frejd, T.; Kihlberg, J .; Magnusson, G.
J . Org. Chem. 1988, 53, 5629.
(33) van der Vleugel, D. J . M.; van Heeswijk, W. A. R.; Vliegenthart,
J . F. G. Carbohydr. Res. 1982, 102, 121.
J . Org. Chem, Vol. 67, No. 21, 2002 7411

Meijer and Ellervik
The amount of unreacted donor was determined from the
NMR spectra of the reaction mixtures, with the TMS signal
from the acceptor/products serving as an internal standard.
Exp er im en ta l Section
NMR spectra were recorded at 300 or 400 MHz. 1H NMR
spectra were assigned using 2D methods (COSY, HETCOR,
long-range HETCOR). Chemical shifts are given in ppm
downfield from the signal for Me4Si, with reference to residual
CHCl3. Reactions were monitored by TLC using alumina plates
coated with silica gel 60 F254 (Merck) and visualized either
by using UV light or by charring with H3PO4 (aqueous 5% dip
solution). Preparative chromatography was performed with
Amicon silica gel (35-70 µm, 60 Å). MeCN and CH2Cl2 used
in glycosylation reactions were stirred overnight with CaH2
and distilled immediately before use. ICl and IBr are com-
mercially available as 1 M solutions in CH2Cl2. The I2 solution
was prepared as follows: molecular iodine was purified by
sublimation and dissolved in CH2Cl2 and MeCN (2:3). Due to
low solubility the concentration was about 0.2 M. Additional
purifications of reaction mixtures were performed on a Sepha-
dex LC-20 column (diameter 1.5 cm, height 1 m) using 1:1 CH2-
Cl2-MeOH as the mobile phase.
Syn th esis of 4-Meth oxyp h en yl (Meth yl(5-a ceta m id o-
4,7,8,9-tetr a -O-a cetyl-3,5-d id eoxy-D-glycer o-R-D-ga la cto-
2-n on u lop yr a n osyl)on a t e)-(1f3)-(4-a zid o-6-O-b en zyl-4-
d eoxy-â-D-ga la ctop yr a n osyl)-(1f4)-[2,3,4-tr i-O-ben zyl-R-
L-fu copyr an osyl]-(1f3)-(2-acetam ido-6-O-ben zyl-2-deoxy-
â-D-glu cop yr a n osid e) (14). Compound 14 was synthesized
according to the representative procedure using the molar
ratios given in Table 1. Compound 14 was obtained in 43%
and 44% yields as a 66:17:17 mixture of the R and â diaster-
eomers together with lactonized product. Unfortunately the â
and lactonized products were inseparable and no characteriza-
tion (and subsequently the anomeric configuration of the
lactone product was not determined) except HRMS was
possible. The R product, however, was obtained as a pure
sample and was characterized as follows.
14-R: [R]20 ) -34.0° (c 0.72, CHCl3); 1H NMR (CDCl3) δ
D
7.15-7.40 (m, 25 H), 6.70-6.92 (m, 4 H), 5.95 (d, 1 H, J ) 7.6
Hz), 5.40-5.47 (m, 1 H), 5.30-5.34 (m, 2 H), 5.20 (d, 1 H, J )
9.7 Hz), 5.13 (d, 1 H, J ) 3.5 Hz), 4.94-5.01 (m, 1 H), 4.95,
4.63 (ABq, 1 H each, J ) 11.6 Hz), 4.87, 4.69 (ABq, 1 H each,
J ) 11.8 Hz), 4.70-4.75 (m, 2 H), 4.58 (d, 1 H, J ) 7.6 Hz),
4.45-4.53 (m, 2 H), 4.35-4.46 (m, 2 H), 4.35-4.40 (m, 1 H),
4.29 (t, 1 H, J ) 8.0 Hz), 4.25 (dd, 1 H, J ) 12.6, 2.7 Hz), 4.17
(dd, 1 H, J ) 10.7, 1.8 Hz), 3.85-4.15 (m, 7 H), 3.66-3.80 (m,
5 H), 3.77 (s, 3 H), 3.74 (s, 3 H), 3.51-3.60 (m, 4 H), 3.11 (s, 1
H), 2.70 (dd, 1 H, J ) 13.2, 4.6 Hz), 2.06-2.14 (m, 1 H), 2.10,
2.10, 2.03, 1.97, 1.90, 1.71 (s, 3 H each), 1.11 (d, 3 H, J ) 6.4
Hz); 13C NMR (CDCl3) δ 171.30, 170.91, 170.71, 170.54, 170.39,
170.23, 168.55 (C-1’’’, J C-1’’’:H-3’’’ax ) 6.9 Hz), 155.56, 151.87,
139.21, 139.15, 139.13, 139.01, 138.19, 128.96, 128.87, 128.83,
128.74, 128.61, 128.57, 128.40, 128.29, 128.22, 127.97, 127.91,
127.69, 127.66, 127.62, 119.16, 114.80, 101.82, 99.60, 98.33,
97.91, 75.32, 75.11, 74.57, 74.36, 74.28, 73.77, 73.25, 73.16,
72.95, 71.21, 70.51, 69.26, 68.77, 68.56, 68.27, 67.37, 67.17,
62.61, 60.95, 56.05, 53.52, 50.06, 38.05, 23.73, 23.63, 21.57,
P r ecip ita tion Exp er im en ts. To a solution of 26 mg of
AgOTf in 2 mL of MeCN was added 0.10 mL of ICl at
room temperature. AgCl precipitation occurred only when a
Lewis base such as pyridine or xanthate donor
1 was
added. The experiment was repeated using MeCN-CH2Cl2 (3:
2) both at room temperature and at -72 °C, with similar
results.
Titr a tion Exp er im en ts. A 1 mL amount of a 0.1 mM
solution of ICl in MeCN was titrated with a 10 mM solution
of AgOTf in MeCN. Spectroscopic data were obtained using a
Varian Cary 300 Bio UV-vis spectrophotometer.
NMR Exp er im en ts. All spectroscopic data were obtained
using a Bruker DRX300 spectrometer at 300 MHz. We chose
to plot the ICl concentration (calculated from [ICl]0 and
∆[MeCl]), which is directly proportional to the ICl2- concentra-
tion.
Figure 2a: to 0.7 mL of CD3CN in an NMR tube was added
4 µL of MeI and 0.13 mL of ICl. Figure 2b: to 0.5 mL of CD3-
CN and 12.8 mg of AgOTf in an NMR tube, protected from
light, was added 5 µL of MeI and 0.10 mL of ICl.
21.28, 21.24, 21.18, 21.11, 17.11; HRMS calcd for C82H97N5O27
-
Na (M + Na) 1606.6269, found 1606.6274.
Rep r esen ta tive P r oced u r e for th e Sia lyla tion Rea c-
tion s. To a stirred solution of 5 (50 mg, 0.122 mmol), 1 (109
mg, 0.183 mmol), and powdered activated molecular sieves (3A,
150 mg) was added CH2Cl2 (1.6 mL) and a solution of AgOTf
(94 mg, 0.367 mmol) in MeCN (2.4 mL). The reaction mixture
was cooled to -72 °C under an argon atmosphere, and a
solution of IBr (1.0 M in CH2Cl2, 0.244 mL, 0.244 mmol) was
added dropwise after 15 min. The reaction mixture was stirred
for 2 h at -72 °C, and then diisopropylamine (0.20 mL, 1.4
mmol) was added and stirring was continued for another 20
min. The reaction mixture was then filtered (SiO2, 25:1-10:1
toluene-EtOH) and concentrated under reduced pressure. The
residue was purified on Sephadex (LH-20, 1:1 CH2Cl2-MeOH)
to give 6 (96 mg, 89%, 8:1 R: â). Compound 6 was acetylated
using standard conditions; the product obtained is identical
with that reported in the literature.12
Rea ction Kin etics. The reactions were carried out as
described in the general procedures, with the following alter-
ations: the reactions were carried out in 5 mL test tubes on a
small scale (0.75 mL of solvent). The IBr solution was added
in a single portion in order to simplify the time measuerement.
The reaction was quenched by addition of a CH2Cl2-cyclohex-
ene-diisopropylamine (2:1:1) solution at -72 °C and then
immediately filtered (SiO2) and concentrated. The residue was
dissolved in CH2Cl2 and the solution was washed twice with
acidified water (H2SO4, pH 2) and once with water and then
concentrated.
14-â: HRMS calcd for C82H97N5O27Na (M + Na) 1606.6269,
found 1606.6274.
14-la cton e: HRMS calcd for C81H93N5O26Na (M + Na)
1574.6006, found 1574.6058.
Meth yl(5-a ceta m id o-4,7,8,9-tetr a -O-a cetyl-3,5-d id eoxy-
D-glycer o-â-D-ga la cto-2-n on u lop yr a n osyl)on a te Br om id e
(7). An anomeric mixture of methyl 5-acetamido-2,4,7,8,9-
penta-O-acetyl-3,5-dideoxy-D-glycero-D-galacto-2-nonulopyra-
nosonate (120 mg, 0.225 mmol) was dissolved in HOAc (2 mL);
Ac2O (1 mL) was added, followed by HBr-HOAc (2 mL, 30%
w/v). The mixture was stirred at room temperature for 1.5 h
and then lyophilized to remove solvent and excess HBr. The
yellowish residue was filtered on a Sephadex LC-20 column
(diameter 1.5 cm, height 10 cm) using acetone as the mobile
phase and then concentrated without heating to yield a
transparent oil consisting of bromide 7 and glycal 4 (3:2). Due
to the instability of bromide 7 (eliminates to glycal 4 upon
storage and handling) the mixture was used immediately in
sialylation reactions.
Ack n ow led gm en t. This work was supported by the
Swedish Natural Science Research Council and the
Knut and Alice Wallenberg Foundation. We thank Dr.
T. Ercegovic for valuable discussions.
Su p p or tin g In for m a tion Ava ila ble: NMR spectrum of
compound 14-R. This material is available free of charge via
the Internet at http://pubs.acs.org.
(34) The counterion (in this case the triflate ion) is capable of
influencing the stereochemical outcome, as shown by Crich: Crich, D.;
Cai, W.; Dai, Z. J . Org. Chem. 2000, 65, 1291.
J O0262412
7412 J . Org. Chem., Vol. 67, No. 21, 2002