1,2-cis-α-Stereoselective Glycosylation Utilizing a Glycosyl-Acceptor-Derived Borinic Ester and Its Application to the Total Synthesis of Natural Glycosphingolipids

Article abstract of DOI:10.1021/acs.orglett.6b02488

1,2-cis-α-Stereoselective glycosylations were conducted using a 1,2-anhydroglucose donor and mono-ol acceptors in the presence of a glycosyl-acceptor-derived borinic ester. Reactions proceeded smoothly under mild conditions to provide the corresponding α-

Full text of DOI:10.1021/acs.orglett.6b02488

Letter
pubs.acs.org/OrgLett
1,2-cis-α-Stereoselective Glycosylation Utilizing a Glycosyl-Acceptor-
Derived Borinic Ester and Its Application to the Total Synthesis of
Natural Glycosphingolipids
Masamichi Tanaka, Daisuke Takahashi,* and Kazunobu Toshima*
Department of Applied Chemistry, Faculty of Science and Technology, Keio University, 3-14-1 Hiyoshi, Kohoku-ku, Yokohama
223-8522, Japan
S
* Supporting Information
ABSTRACT: 1,2-cis-α-Stereoselective glycosylations were conducted using a 1,2-anhydroglucose donor and mono-ol acceptors
in the presence of a glycosyl-acceptor-derived borinic ester. Reactions proceeded smoothly under mild conditions to provide the
corresponding α-glycosides with high stereoselectivity in moderate to high yields. In addition, the present method was applied
successfully to the direct glycosylation of a protected ceramide acceptor and the total synthesis of natural glycosphingolipids
(GSLs).
1,2-cis-α-Glycosidic linkages are prevalent in many biologically
active natural products and glycoconjugates, such as glycolipids.
To understand the precise biological roles and structure−
activity relationships of these carbohydrates, the syntheses of
homogeneous and structurally well-defined carbohydrates have
attracted attention in many research fields.1 However, the
stereoselective synthesis of 1,2-cis-α-glycosides is a challenge
due to the lack of neighboring group participation from a 2-O-
acyl functional group of a glycosyl donor. To overcome this
challenge, efficient indirect2 and direct3 methods have been
developed. For an example of the indirect method, Hindsgaul et
al.4 reported intramolecular aglycon delivery (IAD), which was
extended by Stork et al.,5 Bols,6 and Ogawa and Ito.7 For
examples of the direct method, Liu and Danishefsky8 reported
1,2-cis-α-stereoselective glycosylation of a 1,2-anhydroglucose
donor and stannylated glycosyl acceptors. Boons et al.,9
Turnbull et al.,10 and Fairbanks et al.11 have reported
approaches based on six-membered ring neighboring group
participation. Recently, Demchenko et al.12 reported 1,2-cis-α-
stereoselective glucosylations involving hydrogen bond-medi-
ated aglycon delivery (HAD) with a 4-O-picoloyl thioglucosyl
donor.
corresponding β-mannoside 4 with high stereoselectivity in
high yield, without any additives under mild conditions
(Scheme 1A). Based on these previous studies, the mono-ol
acceptor-derived borinic ester 2 was expected to act as an
activator of 1,2-anhydroglucose 515 to generate the α-glucoside
7 with high stereoselectivity (Scheme 1B). The present report
describes a novel direct and 1,2-cis-α-stereoselective glyco-
sylation of 5 and mono-ol acceptors utilizing a glycosyl-
Scheme 1. (A) β-Stereoselective Mannosylation Using a
Glycosyl-Acceptor-Derived Borinic Ester; (B) α-
Stereoselective Glucosylation using a Glycosyl-Acceptor-
Derived Borinic Ester
Another study reported13 the regio- and 1,2-cis-α-stereo-
selective glycosylation of a 1,2-anhydro donor and a diol
acceptor-derived boronic ester catalyst. However, this method
could be applied only to 1,2- or 1,3-diol acceptors. More
recently, a study14 on the 1,2-cis-β-stereoselective glycosylation
of 3,4,6-tri-O-benzyl (Bn)-1,2-anhydromannose (1) and a
mono-ol acceptor-derived borinic ester 2 was reported. In
that study, the reaction proceeded smoothly to provide the
Received: August 19, 2016
© XXXX American Chemical Society
A
DOI: 10.1021/acs.orglett.6b02488
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
acceptor-derived borinic ester 2 and its application to the total
synthesis of natural GSLs.
Scheme 2. Proposed Catalytic Cycle for Glycosylation of 5
and 9 Using 8c and Proposed Mechanism for the Generation
of 12
Initial investigations led to the selection of 1,2-anhydro-
glucose 5 and diphenylborinic acid (8a),16 di(4-methoxy)-
phenylborinic acid (8b),17 and di(4-fluoro)phenylborinic acid
(8c)17 as the glycosyl donor and arylborinic acids, respectively.
First, glycosylation of 5 and 6-O-benzyl-1-hexanol (9)18 was
attempted using catalytic amounts of 8a−c under different
reaction conditions. The results indicated that the glycosylation
of 5 and 9 using 8a in MeCN at −40 °C for 3 h proceeded to
give glycoside 11 in 60% yield with a 53:47 α/β ratio (Table 1,
Table 1. Glycosylations of 5 and 9 Using Borinic Acids 8a−c
under Several Reaction Conditions
10c increases the activation rate, but reduces the glycosylation
rate.
Taken together, the experimental results and these features
of the borinic esters suggest that the rate-determining step of
this reaction is the activation step, not the intramolecular
glycosylation step, which was the rate-determining step in our
previously reported boronic ester catalyzed 1,2-cis-α-glycosyla-
tion.13
yield (%)
Next, the effect of the reaction solvent on the glycosylation of
5 and 9 was investigated using 8c in toluene, CH2Cl2, Et2O,
and THF. The results showed that the use of toluene and
CH2Cl2 led to β-stereoselectivity (Table 1, entries 4 and 5).
When Et2O was used, the chemical yield of 11 was much lower
than that using MeCN as the solvent (Table 1, entry 6). In
contrast, when THF was used, the reaction proceeded
smoothly to provide 11 in 61% yield with high α-stereo-
selectivity without any accompanying 12 (Table 1, entry 7).
Thus, THF was the best solvent for this reaction. These results
on the effect of the reaction solvent may suggest that strong
coordinating solvents such as THF and MeCN attack from the
β-face of 5 to facilitate the oxonium cation formation.
Next, the reaction time and temperature were optimized.
When glycosylation was conducted at −40 °C for 6 h, the
chemical yield of 11 increased by up to 82% yield but was
accompanied by 12 in 6% yield (Table 1, entry 8). To suppress
overreaction, glycosylation was examined at −60 °C for 6 h,
which resulted in the production of 11 in the best yield (85%)
with complete α-stereoselectivity (Table 1, entry 9).
The scope and limitations of the glycosylation method were
also investigated using several alcohols. The results indicated
that the use of primary alcohols 15−19 and secondary alcohols
20−22 allowed glycosylation to proceed smoothly under mild
conditions to provide the corresponding α-glucosides 25−32 in
high yields with excellent α-stereoselectivity (Table 2, entries
1−8). However, when secondary alcohols 23 and 24 were used,
the chemical yields of the corresponding α-glucosides 33 and
34 were much lower than those of 25−32, probably due to the
low binding affinities of the relatively hindered 23 and 24
toward 8c. Thus, glycosylation of 5 and 23 or 24 in the
presence of 2 equiv of 8c in THF was examined. The
glycosylation proceeded to give 33 and 34 in 59% and 38%
yields, respectively, with complete α-stereoselectivity (Table 2,
entries 9 and 10). These results suggested that the chemical
yield depended on the chemical structure of the glycosyl
acceptor, because the reactivity of the glycosylation step and the
activation step of 5 by the glycosyl-acceptor-derived borinic
ester were strongly influenced by steric effects of the glycosyl
acceptor.
borinic
acid
temp
time
(h)
entry
solvent
(°C)
11 (α:β)
12
1
2
3
4
5
6
7
8
9
8a
8b
8c
8c
8c
8c
8c
8c
8c
MeCN
MeCN
MeCN
toluene
CH2Cl2
Et2O
−40
−40
−40
−40
−40
−40
−40
−40
−60
3
3
3
3
3
3
3
6
6
60 (53:47)
33 (β only)
56 (96:4)
76 (26:74)
52 (23:77)
8 (25:75)
61 (95:5)
82 (95:5)
85 (α only)
0
0
17
0
0
0
THF
0
THF
6
THF
0
entry 1). To improve the α-stereoselectivity, the electrostatic
effect of substituents on the benzene ring in the borinic esters
was investigated using 8b and 8c. When 8b, which possesses an
electron-donating methoxy group, was used, complete β-
stereoselectivity was obtained (Table 1, entry 2). In sharp
contrast, when 8c, which possesses an electron-withdrawing
fluorine group, was used, glycoside 11 was obtained in 56%
yield with high α-stereoselectivity along with disaccharide 12 in
17% yield as a byproduct (Table 1, entry 3). The configurations
1
of both glycosidic bonds in 12 were confirmed to be α by H
NMR analysis. This result suggests that the borinic ester 10c
induced sequential α-stereoselective glycosylation to provide 12
(Scheme 2). In addition, the results of the electrostatic effect
studies suggested that the electron-donating group in 10b
reduced the Lewis acidity of the boron atom, resulting not in
oxonium cation formation but SN2 intermolecular nucleophilic
attack of 9 from the β-face of 5. In contrast, the electron-
withdrawing group in 10c increased the Lewis acidity of the
boron atom, resulting in oxonium cation formation, which led
to SN1 type intramolecular nucleophilic attack of the boron-
bound oxygen atom in 13c from the same α-face of 5.
Furthermore, according to the chemical features of the borinic
esters, it is reasonable to assume that the electron-donating
group in 10b reduces the activation rate, but increases the
glycosylation rate, whereas the electron-withdrawing group in
B
DOI: 10.1021/acs.orglett.6b02488
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
and GSL-1′ using azido-sphingosine and N-Boc-sphingosine,
respectively. In contrast, although Seeberger et al.28 reported
the total synthesis of GSL-1′ using a protected ceramide
acceptor, the α-stereoselectivity probably could be improved.
The present method was designed to overcome both of these
obstacles since the protected ceramide-derived borinate ester
could increase the nucleophilicity of the boron bound oxygen
atom and lead to 1,2-cis-α-stereoselective glycosylation (Figure
1C). The synthetic schemes of GSL-1 and GSL-1′ are
summarized in Scheme 3.
Table 2. Glycosylations of 5 and Alcohols Using a Catalytic
Amount of 8c
Scheme 3. Total Synthesis of GSL-1 and GSL-1′
entry acceptor temp (°C) time (h) product yield (%) α/β ratio
1
2
3
4
5
6
7
8
9
15
16
17
18
19
20
21
22
23
24
−80
−60
−80
−80
−40
−40
−40
−40
−80
−40
6
6
25
26
27
28
29
30
31
32
33
34
99
85
99
89
99
75
82
80
59
38
α only
α only
α only
α only
α only
α only
α only
α only
α only
α only
3
3
6
6
a
b
6
24
6
ab
,
10
24
a
b
MeCN was employed as a solvent. 2 equiv of 8c were used.
In addition, the present glycosylation method was applied to
the total synthesis of biologically active natural GSLs, GSL-1,
and GSL-1′ (Figure 1A). GSL-1 was isolated from F. devorans
Figure 1. Nucleophilicity of the C-1 OH group in the ceramide
acceptor.
ATCC 1082919 and S. paucimobilis,20 and GSL-1′ was isolated
from S. wittichii21 and S. yanoikuyae.22 Structurally, these
compounds contain the hydrophobic moiety ceramide, and the
hydrophilic monosaccharide group 1,2-cis-α-glucuronic or
galacturonic acid. These GSLs are recognized by human
natural killer T (NKT) cells after binding to CD1d and induce
a significant immune response.23 From a synthetic standpoint,
two problems need to be overcome in the direct glycosylation
of a glycosyl donor and ceramide acceptor: 1,2-cis-α-stereo-
selectivity and low nucleophilicity of the C1 hydroxyl group in
the ceramide acceptor due to hydrogen bond formation
between the electron pair on the C1 oxygen and amide
hydrogen (Figure 1B).24 To manage the low nucleophilicity,
sphingosine-derived acceptors, such as an azido-sphingosine,25
have been widely used in the synthesis of various GSLs. Savage
et al.26 and Wong et al.27 reported the total synthesis of GSL-1
Initially, alcohol 3628 was prepared from 1-tetradecene (35).
Protection of a hydroxyl group in 36 as a p-methoxybenzyl
(PMB) ether, followed by deprotection of a Tr group, provided
37 in 77% overall yield. Oxidation of the resulting primary
hydroxyl group in 37, followed by amidation with 39,29
afforded the ceramide acceptor 40 in good yield. Next,
glycosylation of 40 and 41 (Scheme S1 in the Supporting
Information (SI)) was performed using a catalytic amount of
8c. The result showed that the desired α-glucoside 42 was
obtained in 80% yield with excellent α-stereoselectivity.
Deprotection of the TBS group in 42 and oxidation of the
resulting primary hydroxyl group, followed by removal of the
1
benzyl and PMB groups, furnished the desired GSL-1. The H
C
DOI: 10.1021/acs.orglett.6b02488
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
NMR, 13C NMR, and HRMS data for a sample of synthetic
GSL-1 were identical to those reported.26
(7) Ito, Y.; Ogawa, T. Angew. Chem., Int. Ed. Engl. 1994, 33, 1765−
1767.
(8) Liu, K. K.-C.; Danishefsky, S. J. J. Org. Chem. 1994, 59, 1895−
1897.
(9) (a) Kim, J.-H.; Boons, G.-J. Angew. Chem., Int. Ed. 2005, 44, 947−
949. (b) Kim, J.-H.; Yang, H.; Park, J.; Boons, G.-J. J. Am. Chem. Soc.
2005, 127, 12090−12097.
(10) Fascione, M. A.; Adshead, S. J.; Stalford, S. A.; Kilner, C. A.;
Leach, A. G.; Turnbull, W. B. Chem. Commun. 2009, 5841−5843.
(11) Cox, D. J.; Fairbanks, A. J. Tetrahedron: Asymmetry 2009, 20,
773−780.
(12) Yasomanee, J. P.; Demchenko, A. V. J. Am. Chem. Soc. 2012,
134, 20097−20102.
Next, glycosylation of 40 and 43 (Scheme S2 in SI) was
examined using a catalytic amount of 8c. This reaction also
proceeded smoothly to provide the desired α-galactoside 44 in
84% yield with excellent α-stereoselectivity. Deprotection of the
TBS group in 44, followed by oxidation of the resulting primary
hydroxyl group, gave partly protected GSL-1′ 45. Removal of
the Bn and PMB groups in 45 furnished GSL-1′. Finally,
methyl esterification of 45, followed by removal of the Bn and
PMB groups, furnished the GSL-1′ methyl ester, confirmed by
22
1
comparing the H NMR data to those reported.
(13) Nakagawa, A.; Tanaka, M.; Hanamura, S.; Takahashi, D.;
Toshima, K. Angew. Chem., Int. Ed. 2015, 54, 10935−10939.
(14) Tanaka, M.; Nashida, J.; Takahashi, D.; Toshima, K. Org. Lett.
2016, 18, 2288−2291.
In conclusion, direct and highly α-stereoselective glycosyla-
tion of 1,2-anhydroglucose 5 and mono-ol acceptors has been
developed utilizing a glycosyl-acceptor-derived borinic ester
under mild conditions. The use of di(4-fluoro)phenylborinic
acid (8c) in THF or MeCN was effective for glycosylation with
several mono-ol acceptors. This method was applied success-
fully to the direct glycosylation of a protected ceramide
acceptor and the total synthesis of GSL-1 and GSL-1′. Detailed
mechanistic studies of this method, its application to other
types of acceptors, and the synthetic studies of other
biologically active compounds using this method are underway.
(15) Halcomb, R. L.; Danishefsky, S. J. J. Am. Chem. Soc. 1989, 111,
6661−6666.
(16) Lee, D.; Newman, S. G.; Taylor, M. S. Org. Lett. 2009, 11,
5486−5489.
(17) Mori, Y.; Kobayashi, J.; Manabe, K.; Kobayashi, S. Tetrahedron
2002, 58, 8263−8268.
(18) Yadav, J. S.; Subba Reddy, U. V.; Subba Reddy, B. V.
Tetrahedron Lett. 2009, 50, 5984−5986.
(19) Yamamoto, A.; Yano, I.; Masui, M.; Yabuuchi, E. J. Biochem.
1978, 83, 1213−1216.
ASSOCIATED CONTENT
■
(20) Kawahara, K.; Seydel, U.; Matsuura, M.; Danbara, H.; Rietschel,
S
E. T.; Zahringer, U. FEBS Lett. 1991, 292, 107−110.
̈
* Supporting Information
(21) Kawahara, K.; Kubota, M.; Sato, N.; Tsuge, K.; Seto, Y. FEMS
Microbiol. Lett. 2002, 214, 289−294.
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.or-
glett.6b02488.
(22) Kawahara, K.; Sato, N.; Tsuge, K.; Seto, Y. Microbiol. Immunol.
2006, 50, 67−71.
1
Experimental methods and synthetic details; H and 13C
(23) Kinjo, Y.; Wu, D.; Kim, G.; Xing, G.-W.; Poles, M. A.; Ho, D.
D.; Tsuji, M.; Kawahara, K.; Wong, C.-H.; Kronenberg, M. Nature
2005, 434, 520−525.
NMR data (PDF)
(24) (a) Morales-Serna, J. A.; Boutureira, O.; Díaz, Y.; Matheu, M. I.;
AUTHOR INFORMATION
Castillon, S. Carbohydr. Res. 2007, 342, 1595−1612. (b) Di Benedetto,
́
■
R. D.; Zanetti, L.; Varese, M.; Rajabi, M.; Di Brisco, R. D.; Panza, L.
Org. Lett. 2014, 16, 952−955.
Corresponding Authors
*E-mail: dtak@applc.keio.ac.jp.
*E-mail: toshima@applc.keio.ac.jp.
Notes
(25) Schmidt, R. R.; Zimmermann, P. Angew. Chem., Int. Ed. Engl.
1986, 25, 725−726.
(26) Long, X.; Deng, S.; Mattner, J.; Zang, Z.; Zhou, D.; McNary, N.;
Goff, R. D.; Teyton, L.; Bendelac, A.; Savage, P. B. Nat. Chem. Biol.
2007, 3, 559−564.
The authors declare no competing financial interest.
(27) Wu, D.; Xing, G.-W.; Poles, M. A.; Horowitz, A.; Kinjo, Y.;
Sullivan, B.; Bodmer-Narkevitch, V.; Plettenburg, O.; Kronenberg, M.;
Tsuji, M.; Ho, D. D.; Wong, C.-H. Proc. Natl. Acad. Sci. U. S. A. 2005,
102, 1351−1356.
ACKNOWLEDGMENTS
■
This research was supported in part by the MEXT-supported
Program for the Strategic Research Foundation at Private
Universities, 2012-2016, and JSPS KAKENHI Grant Numbers
JP16H01161 in Middle Molecular Strategy and JP16K05781.
(28) Stallforth, P.; Adibekian, A.; Seeberger, P. H. Org. Lett. 2008, 10,
1573−1576.
(29) Ndonye, R. M.; Izmirian, D. P.; Dunn, M. F.; Yu, K. O. A.;
Porcelli, S. A.; Khurana, A.; Kronenberg, M.; Richardson, S. K.;
Howell, A. R. J. Org. Chem. 2005, 70, 10260−10270.
REFERENCES
■
(1) (a) Carbohydrates in Chemistry and Biology; Ernst, B., Hart, G.
W., Sinay, P., Eds; Wiley-VCH: Weinheim, 2000; Vols. 1−4. (b)
̈
Glycoscience, Chemistry and Chemical Biology; Fraser-Reid, B. O.,
Tatsuta, K., Thiem, J., Eds.; Springer: Berlin, 2001.
(2) For selected reviews, see: (a) Jung, K.-H.; Muller, M.; Schmidt, R.
̈
R. Chem. Rev. 2000, 100, 4423−4442. (b) Cumpstey, I. Carbohydr. Res.
2008, 343, 1553−1573. (c) Ishiwata, A.; Lee, Y. J.; Ito, Y. Org. Biomol.
Chem. 2010, 8, 3596−3608.
(3) For selected reviews, see: (a) Crich, D. J. Org. Chem. 2011, 76,
9193−9209. (b) Nigudkar, S. S.; Demchenko, A. V. Chem. Sci. 2015, 6,
2687−2704.
(4) Barresi, F.; Hindsgaul, O. J. Am. Chem. Soc. 1991, 113, 9376−
9377.
(5) Stork, G.; Kim, G. J. Am. Chem. Soc. 1992, 114, 1087−1088.
(6) Bols, M. J. Chem. Soc., Chem. Commun. 1992, 913−914.
D
DOI: 10.1021/acs.orglett.6b02488
Org. Lett. XXXX, XXX, XXX−XXX