α-selective glycosylation with 5-thioglucopyranosyl donors; Synthesis of an isomaltotetraoside mimic composed of 5-thioglucopyranose units

Article abstract of DOI:10.1016/S0960-894X(03)00072-6

A general method for α-selective glycosylation with 5-thioglucopyranosyl donors followed by efficient deprotection of the resulting products was developed. This methodology was utilized in the synthesis of an isomaltotetraoside analogue.

Full text of DOI:10.1016/S0960-894X(03)00072-6

Bioorganic & Medicinal Chemistry Letters 13 (2003) 1063–1066
ꢀ-Selective Glycosylation with 5-Thioglucopyranosyl Donors;
Synthesis of an IsoMaltotetraoside Mimic Composed of
5-Thioglucopyranose Units
Hiroko Matsuda,b Keiichiro Ohara,a Yasuharu Morii,a Masaru Hashimoto,a,b,
Kazuo Miyairia,b and Toshikatsu Okunoa,b
*
aFaculty of Agriculture and Life Science, Hirosaki University, Bunkyo-Cho 3, Hirosaki 036-8561, Japan
bThe United Graduate School of Agricultural Sciences, Iwate University, Bunkyo-Cho 3, Hirosaki 036-8561, Japan
Received 7 November 2002;accepted 14 January 2003
Abstract—A general method for a-selective glycosylation with 5-thioglucopyranosyl donors followed by efficient deprotection of
the resulting products was developed. This methodology was utilized in the synthesis of an isomaltotetraoside analogue.
# 2003 Elsevier Science Ltd. All rights reserved.
Carbomimetics, analogues which closely mimic carbo-
hydrate structure, are predicted to be excellent probes
not only for the investigation of the details of enzymatic
reaction mechanisms that involve specific protein–sugar
interactions, but also as novel drugs for some digestive
or infective diseases.1,2 Many saccharide analogues have
been reported3ꢀ5 and most of these have focused on exo-
type glycosidases that hydrolyze the terminal glycosidic
bond through recognition of the terminal carbohydrate
subsite.6 However, in the case of endo-type glycosidases
that cleave internal glycosidic linkages, the development
of enzyme inhibitors requires that oligomeric structures
be prepared since endo-glycosidases utilize more widely
dispersed subsites for recognition. Due to the increased
complexity of substrate binding for endo-glycosidases,
an inhibitor of these enzymes generally must incorpo-
rate additional functionality beyond that present at the
cleavage site. Thus, mechanistic studies of endo-glyco-
sidases are relatively less advanced than those on exo-
glycosidases due to an absence of effective inhibitors of the
endo-type enzymes.7 Recently, oligosaccharide mimics
containing C-glycoside linkages have been developed as
endo-glycosidase inhibitors.8 Since the interaction
between amino acid residues and glycosidic oxygens in
the enzyme–substrate complex must be considered, we
anticipated that sulfur substituted saccharides would
provide promising inhibitors of endo-glycosidases.
However, the increased length of the carbon–sulfur
bond compared to that of a carbon–oxygen bond could
lead to a significantly altered molecular display of the
atoms directly involved in enzyme binding. The oligo-
merization of a series of sulfur-substituted mono-
saccharides might result in substantial alteration in
molecular size when the acetalic oxygen is replaced with
sulfur. For example, as shown in Figure 1, molecular
Figure 1. Comparison of a ‘natural’ oligosaccharide model with mod-
els of analogues containing thioglycosides and thiopyranoses.
*Corresponding author. Fax: +81-172-39-3782;e-mail: hmasaru@
cc.hirosaki-u.ac.jp
0960-894X/03/$ - see front matter # 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0960-894X(03)00072-6

1064
H. Matsuda et al. / Bioorg. Med. Chem. Lett. 13 (2003) 1063–1066
modeling calculations indicate that a tetramer of an
a-(1!4)-glycan with thioglycosidic linkages (‘thioglyco-
side’ model) results in a more extended molecular
structure as compared with that of a ‘natural’ tetra-
saccharide.9 Endo-glycosidases may no longer be capa-
ble of binding this type of analogue. However, as
supported by similar molecular modeling studies, a tetra-
mer of 1,4-linked thiopyranoses (‘thiopyranose’ model,
Fig. 1) may better absorb the deviations of bond lengths
and bond angles between ethers and thioethers, thereby
minimizing structural change. Accordingly, oligothio-
saccharides may be potential inhibitors of endo-glyco-
sidases. In this report, we describe a general method for
the a-selective glycosylation with 5-thioglucopyranosyl
donors and an efficient deprotection protocol of the
polysaccharide products. Utilizing this methodology,
the synthesis of sulfur-substituted isomaltotetraoside 1
is also described.
Donors 3 and 4 were prepared by standard method-
ology as shown in Scheme 1. It was found that
O-benzylprotected donor 3 gave glycosides 14 and 15
in good yields by treating 3 with acceptors 8 and 9,
respectively, along with
a
catalytic amount of
TESOTf (0.05–0.1 equiv) at ꢀ78 ꢁC. The 1H NMR
spectra of 14 and 15 revealed that both glycosyla-
tions proceeded in high a-stereoselectivity (>98:2).
Similarly, O-(4-methoxyphenyl)methyl (MPM) pro-
tected donor 4 provided adducts 16–19 in a highly
a-stereoselective manner even though the yields were
slightly lower than that of the reactions with donor
3. In these last examples, a small amount of rela-
tively polar material was observed by TLC analysis,
suggesting that some MPM groups were cleaved dur-
ing the reactions.
Table 1. Glycosylation with 5-thioglucopyranosyl donors
Run
1
Donors
Acceptors
Disaccharides (%)
a/ba
2
12 (5%)
>95:5b
2
2
13 (10%)
>95:5b
Several groups have reported extensive efforts directed
at glycosylation reactions with 5-thiopyranosyl donors.
Those reports typically employed O-acetyl-protected
imidate 2 as the glycosyl donor with various promoters.
However, in these examples, both the yields and the
stereoselectivity of the reactions were dependent on the
glycosyl acceptor.10ꢀ13 For example, Mehta et al. dis-
closed the glycosylation in 80% yield but with low
selectivity (a/b=1.5:1),11 while Izumi et al. obtained the
a-disaccharide exclusively in only 22% yield.12 In our
preliminary experiments, we also examined the glyco-
sylation using donor 2, but the yields were quite low
under many different reaction conditions (e.g., Run 1
and 2, Table 1). A general and highly stereoselective
glycosylation reaction was required for the synthesis of
oligothiosaccharides. Thus, we first investigated a gen-
eral glycosylation reaction employing thiogluco-
pyranosyl donors 3 and 4.
3
4
5
6
3
3
4
4
8
9
8
9
14 (94%)
15 (81%)
16 (60%)
17 (94%)
>98:2b
>98:2b
94:6
>98:2b
7
8
4
4
18 (87%)
>98:2b
19 (82%)
88:12
aThe ratios were estimated by 400 MHz H NMR spectra.
bThe b-isomers were not detected by 400 MHz H NMR spectra.
1
1
Scheme 1. (a) DHP, p-TsOH, CH2Cl2 (95%);(b) NaOMe, MeOH
(87%);(c) BnBr (88% for 3) or MPMBr (95% for 4), NaH, DMF;
(d) H+, MeOH (90% for 3, 81% for 4);(e) CCl 3CN, DBU, CH2Cl2,
0 ꢁC (57% for 3, 74% for 4).

H. Matsuda et al. / Bioorg. Med. Chem. Lett. 13 (2003) 1063–1066
1065
methyl 50-thio-a-isomaltoside (21) in 77 and 88% yield,
respectively, over two steps.16 In each of these cases, the
hydroquinone derivative could be separated easily by
column chromatography because of the increased
polarity of the products. Oxidation of the thioacetal
moiety with DDQ was not observed. This was con-
firmed by observing the molecular ion peaks of both 20
and 21 at m/z=371 ([MꢀH]ꢀ) by FAB-MS (negative
mode). In contrast, the same operations for 16 and 17
resulted only in decomposition of the substrates and/or
products, suggesting that the C1 position of the thio-
glucose moiety must be protected for DDQ treatment.
Scheme 2. (a) NaOMe, MeOH;(b) DDQ, aq CH 2Cl2.
With effective glycosylation and deprotection protocols
in hand, a synthesis of isomaltotetraose mimic 1 was
performed (Scheme 3). Glycosyl donor 25 was prepared
from 512 by sequential (1) C1–OH protection as the tet-
rahydropyranyl (THP) (!7) or ethoxyethyl (EE) ethers
(!22), (2) removal of all acetyl groups, (3) selective tri-
tyl ether formation of the primary alcohol, (4) protec-
tion of the remaining hydroxyl groups as MPM ethers
(!23, 24), (5) regeneration of the C1–OH, and (6) for-
mation of the trichloroacetimidate17 (!25). Acidic
removal the THP group of 23 gave the alcohol in low
yield (30%) due to simultaneous cleavage of the trityl
ether. It was found that the more labile EE group of 24
was deprotected selectively by treatment with pyr-
idinium p-toluenesulfonate (PPTS) in 92% yield.18
With disaccharides 14–19 in hand, deprotection reactions
were next attempted. Although glycosylation employing
donor 3 was effective, hydrogenolysis of the O-benzyl
ethers in either 14 or 15 was unsuccessful, providing a
number of partially deprotected alcohols. The sluggish-
ness of this reaction was most likely due to poisoning of
the palladium catalyst by the sulfide functionality.11,14
Next, alternative deprotection methods were examined
using MPM-masked disaccharides 16–19. Treatment
with excess 2,3-dichloro-5,6-dicyano-1,4-benzoquinone
15
(DDQ) in aqueous CH2Cl2 appeared to smoothly
deprotect the MPM group based on TLC analysis.
However, the resulting alcohols could not be separated
by silica gel chromotography from the corresponding
hydroquinone that was generated in large quantity dur-
ing the reaction. Acetylation of the crude reaction pro-
ducts (Ac2O in pyridine) resulted in a complex mixture.
Deprotection of disaccharides 18 and 19 could be
achieved by changing the order of the deprotection
reactions (Scheme 2). After the benzoyl groups were
removed by treatment with sodium methoxide in
methanol, the resulting products were further depro-
tected with DDQ. Efficient cleavage of all of the MPM
groups provided methyl 50-thio-a-maltoside (20) and
As expected, the glycosylation reaction of acceptor 11
with donor 25 proceeded smoothly, providing, a-di-
saccharide 26 in 67% yield along with a small amount
of the b-isomer (7%), after chromatographic purifica-
tion. After selective cleavage of the trityl ether under
acidic conditions, the resulting alcohol 27 was subjected
to a second glycosylation to afford the trisaccharide 28.
This material was further converted to tetrasaccharide
30 by similar treatment. Interestingly, formation of
b-glycosides was not observed in the latter two glycosy-
1
lation steps based on H NMR spectroscopy. Finally,
Scheme 3. (a) For 22;ethyl vinyl ether, PPTS, CH Cl2, 0 ꢁC (96%), then NaOMe, MeOH (86%);(b) TrCl, Py, DMAP (31% for 23, 81% for 24);
(c) MPMBr, NaH, DMF (69% for 23, 62% for 24);(d) PPTS, MeOH (30% from 23), or PPTS, EtOH–PrOH (92% from 24);(e) CCl CN, DBU,
2
3
CH2Cl2, 0 ꢁC (81%);(f) 11, TESOTf, MS4A, CH2Cl2, ꢀ78 ꢁC (74%, a/b=10:1);(g) p-TsOH, MeOH–THF (71% for 27, 67% for 29);(h) 25,
TESOTf, MS4A, CH2Cl2, ꢀ78 ꢁC (47%);(i) 4, TESOTf, MS4A, CH2Cl2, ꢀ78 ꢁC (67%);(j) NaOMe, MeOH;(k) DDQ, aq CH 2Cl2 (65%, two
steps).

1066
H. Matsuda et al. / Bioorg. Med. Chem. Lett. 13 (2003) 1063–1066
all protective groups of tetramer 30 were successfully
removed by employing the methodology described
above (NaOMe in methanol, followed by DDQ in aqu-
eous CH2Cl2), giving free sulfur substituted methyl iso-
maltotetraoside 1 in 65% yield (two steps).19 The
tetrameric structure of 1 was confirmed by the observa-
tion of four signals of equatorial oriented anomeric
protons at d=4.79, 4.82, 4.83, and 4.88 ppm (J=2.9,
13. Hashimoto, H.;Izumi, M. Tetrahedron Lett. 1993, 34,
4949.
14. Deprotection under Birch’s conditions was not examined.
15. Oikawa, Y.;Yoshioka, T.;Yonemitsu, O.
Tetrahedron
Lett. 1982, 23, 885.
16. Spectral data, 20: 1H NMR (400 MHz, D2O, 38 ꢁC) d 3.05
(1H, m, C05H), 3.42 (3H, s, C1OCH3), 3.59 (1H, dd, J=3.9,
9.7 Hz, C2H), 3.62–3.67 (2H, C03H, C04H), 3.73–3.77 (2H,
C4H, C5H), 3.82–3.85 (2H, C02H, C6H), 3.88–3.90 (3H, C06H
ꢂ2, C6H), 3.95 (1H, t, J=9.7 Hz, C3H), 4.81 (1H, d,
J=3.9 Hz, C1H), 5.33 (1H, d, J=3.4 Hz, C01H). 13C NMR
(100 MHz, D2O, 38 ꢁC) d 43.9 (C05), 55.2 (C1OCH3), 60.2
(C06), 61.0 (C6), 70.2 (C5), 71.3 (C2), 73.4 (C04), 73.9 (C03),
74.1 (C3), 75.5 (C2), 75.6 (C4), 82.8 (C01), 99.2 (C1). FAB-MS
(%, rel. int.) 371 ([MꢀH]ꢀ, 100): HR-MS (FAB, negative)
calcd for C13H23O10S [MꢀH]ꢀ: 371.1012, found m/z
1
2.9, 2.9, and 3.9 Hz, respectively) in the H NMR spec-
trum as well as the molecular ion peak at m/z=727
([MꢀH]ꢀ) by FAB-MS (negative mode). Further, the
13C NMR spectrum (25 carbons) fully supported the
structure of 1.
As described in this report, we have developed a general
a-stereoselective glycosylation employing 5-thiogluco-
pyranoses as well as a deprotection protocol that is
applicable to the resulting oligothiosaccharides. Studies
on the enzyme binding properties of 1 and the synthesis
of additional oligothiosaccharides are under investi-
gation in our laboratory.
=371.0975. 21: H NMR (400 MHz, D2O, 38 ꢁC) d 3.10 (1H,
1
ddd, J=3.4, 4.9, 10.3 Hz, C05H), 3.43 (3H, s, C1OCH3), 3.49
(1H, dd, J=9.0, 10.0 Hz, C4H), 3.57 (1H, dd, J=3.9, 9.7 Hz,
C2H), 3.64 (1H, dd, J=8.8, 10.2 Hz, C04H), 3.67 (1H, dd,
J=9.0, 9.7 Hz, C3H), 3.69 (1H, dd, J=8.8, 9.3 Hz, C03H), 3.72
(1H, dd, J=2.0, 11.0 Hz, C6H), 3.83 (1H, m, C5H), 3.85 (1H,
dd, J=2.9, 9.3 Hz, C02H), 3.87 (1H, dd, J=3.4, 11.4 Hz,
C06H), 3.92 (1H, dd, J=4.9, 11.4 Hz, C06H), 4.13 (1H, dd,
J=4.6, 11.2 Hz, C6H), 4.77 (1H, d, J=3.0 Hz, C01H), 4.82
(1H, d, J=3.9 Hz, C1H). 13C NMR (100 MHz, D2O, 38 ꢁC) d
43.1 (C05), 55.3 (C1OCH3), 60.2 (C06), 66.6 (C6), 69.7 (C4),
70.2 (C5), 71.3 (C2), 73.5 (C04), 73.6 (C3), 74.2 (C03), 75.4
(C02), 81.9 (C01), 99.4 (C1). FAB-MS (%, rel. int.) 371
([MꢀH]ꢀ, 21), HR-MS (FAB, negative) calcd for C13H23O10S
[MꢀH]ꢀ: 371.1012, found m/z 371.1000.
Acknowledgements
We would like to thank Professor Jun Kawabata and
Dr. Eri Fukushi of Hokkaido University for measure-
ments of mass spectra and for fruitful discussions. We
are grateful to Dr. Craig A. Parish of Merck Research
Laboratories for kindly reading and correcting this
manuscript.
17. Schmidt, R. R. Angew. Chem., Int. Ed. Engl. 1986, 25,
212.
18. Deprotection of EE group in 24 was performed by
sequential operations of (i) stirring in EtOH in the pre-
sence of PPTS for 30 min, (ii) evaporation of the mixture,
and (iii) stirring for 2 h after dilution with PrOH. Sub-
strate 24 was insoluble in EtOH. The reactions in PrOH
or in a mixed solvent system (PrOH:EtOH=3:1) were less
effective.
19. A solution of 30 (11.5 mg, 5.10 mmol) in MeOH (2.0 mL)
was stirred with NaOMe (2.5 mg, 46.3 mmol) at room tem-
perature for 1 h. The mixture was neutralized by the addition
of DOWEX 50W (H+ form), filtered, and concentrated in
vacuo. Without further purification, the residue was stirred
with DDQ (46.2 mg, 204 mmol) in a mixture of CH2Cl2
(1.0 mL) and H2O (100 mL) at room temperature for 12 h.
After concentration, the residue was purified by silica gel col-
References and Notes
1. Glycoscience; Synthesis of Substrate Analogues and Mimet-
ics. Driguez, H., Thiem, J. Eds. Springer: Berlin, 1999.
2. Glycoscience; Synthesis of Oligosaccharides and Glyco-
conjugates. Driguez, H., Thiem, J. Eds. Springer: Berlin, 1999.
3. Witczak, Z. J. Curr. Med. Chem. 1999, 6, 165.
4. Spohr, U.;Bach, M.;Spiro, R. G. Can. J. Chem. 1993, 71,
1919.
5. Gurjar, M. K.;Nagaprasad, R.;Ramana, C. V.
hedron Lett. 2002, 43, 7577.
Tetra-
6. Davies, G. J.;Wilson, K. S.;Henrissat, B.
1997, 321, 557.
7. Shimizu, T.;Nakatsu, T.;Miyairi, K.;Okuno, T.;Kato, H.
Biochemistry 2002, 41, 6651.
8. Dondoni, A.;Marra, A.;Mizuno, M.;Ciovannini, P. P. J.
Org. Chem. 2002, 67, 4186.
9. PC Spartan Pro version 1.0.8 by wavefunction Ltd.Co.
10. Yuasa, H.;Hashimoto, H. Rev. Heteroatom Chem. 1999,
19, 35.
11. Mehta, S.;Jordan, K. L.;Weimar, T.;Kreis, U. C.;
Batchelor, R. J.;Einstein, F. W. B.;Pinto, B. M. Tetrahedron:
Asymmetry 1994, 5, 2367.
12. Izumi, M.;Suhara, Y.;Ichikawa, Y. J. Org. Chem. 1998,
63, 4811.
Biochem. J.
umn chromatography (MeOH/AcOEt=40:60) to give 1
(2.4 mg, 65% in two steps). 1H NMR (400 MHz, D2O, 43 ꢁC) d
3.24 (1H, brddd, J=3.4, 4.8, 9.8 Hz), 3.32–3.37 (1H, m), 3.39–
3.44 (1H, m), 3.50 (3H, s), 3.56 (1H, t, J=9.8 Hz), 3.64 (1H,
dd, J=3.9, 9.8 Hz), 3.67–3.79 (7H), 3.87–4.00 (9H), 4.10–4.17
(2H), 4.22 (1H, dd, J=4.9, 9.8 Hz), 4.79 (1H, d, J=2.9 Hz),
4.82 (1H, d, J=2.9 Hz), 4.83 (1H, d, J=2.9 Hz), 4.88 (1H, d,
J=3.9 Hz). 13C NMR (100 MHz, D2O) d 40.8 (d), 40.9 (d),
43.3 (d), 55.4 (q), 66.7 (t), 60.3 (t), 66.9 (t), 67.0 (t), 69.8 (d),
70.2 (d), 71.4 (d), 73.6 (d), 73.7 (d), 74.23 (d), 74.25 (d), 74.31
(d), 74.32 (d), 74.34 (d), 75.28 (d), 75.34 (d), 75.40 (d), 81.68
(d), 81.71 (d), 81.92 (d), 99.5 (d). FAB-MS (%, rel. int.) 727
([MꢀH]ꢀ, 19), HR-MS (FAB, negative) calcd for C25H43O18S3
[MꢀH]ꢀ: 727.1612, found m/z 727.1591.