Reaction of thiophenol with glucal epoxides: X-ray structure of 3,4,6-tri-O-tert-butyldimethylsilyl-1-S-phenyl-1-thio-α-D-glucopyranoside

Article abstract of DOI:10.1039/a704534d

The reaction of the tert-butyldimethylsilyl protected glucal epoxide 1 with thiophenol under a variety of conditions gives three distinct adducts, including the α-phenylthioglycoside 2a, which exists both in solution and in the solid state in ¹

Full text of DOI:10.1039/a704534d

View Article Online / Journal Homepage / Table of Contents for this issue
Reaction of thiophenol with glucal epoxides: X-ray structure of
3,4,6-tri-O-tert-butyldimethylsilyl-1-S-phenyl-1-thio-a-d-glucopyranoside
Caroline Walford, Richard F. W. Jackson,* Nicholas H. Rees, William Clegg and Sarah L. Heath
Department of Chemistry, Bedson Building, The University of Newcastle, Newcastle upon Tyne, UK NE1 7RU
O
SOnPh
OH
The reaction of the tert-butyldimethylsilyl protected glucal
O
ButMe2SiO
ButMe2SiO
ButMe2SiO
epoxide 1 with thiophenol under a variety of conditions gives
three distinct adducts, including the a-phenylthioglycoside
2a, which exists both in solution and in the solid state in a 1C4
conformation in which the C-2, C-3, C-4 and C-5 sub-
stituents are all axial.
O
i
+
ButMe2SiO
2b
OSiMe2But
OSiMe2But
1
2a n = 0
ii
5
n = 2
Scheme 2 Reagents and conditions: i, PhSH, 0 °C to room temp., THF, 16
h; ii, MCPBA, NaHCO3, CH2Cl2, 0 °C, 1.5 h
1-Phenylthioglucosides have been widely used in glycosylation
reactions. In the context of a project to develop nucleophilic
glycosyl donors, we were interested in the development of
effective routes to protected a- and b-phenylthioglucosides 2a
and 2b. Our planned approach involved the reaction of the
epoxide 1 (prepared by oxidation of the readily available glucal
with dimethyldioxirane)1 with phenylthiolate, a reaction which
had been previously reported on the corresponding tribenzyl
protected derivative by Spilling (using NaSPh)2 and Da-
nishefsky (using Bu4NSPh).3 Spilling obtained a low yield, and
Danishefsky obtained a mixture of both a- and b-phenyl-
thioglucosides. When the epoxide 1 was treated with lithium
phenylthiolate (Scheme 1), the adduct 3 was cleanly obtained
(72%), in which the initial product of ring-opening had
undergone a silyl shift. However, treatment of the epoxide 1
with thiophenol in the presence of triethylamine did give the
expected sulfide 2b (85%). The structures of the two products
were assigned on the basis of extensive two-dimensional NMR
studies, including the preparation of acetylated derivatives.
While conversion of the sulfide 2b into the tetra-silylated
derivative 4 (tert-butyldimethylsilyl trifluoromethanesulfonate,
2,6-lutidine, 278 °C to room temp., 16 h) proceeded reasonably
efficiently (58%), the analogous reaction on 3 was very slow,
and gave 4 in very poor yield (12%), testifying to the hindered
nature of the C-3 hydroxy group.4 We believe that the best
explanation for these results is that the presence of two bulky
protecting groups at C-3 and C-4 is unfavourable, and it is the
relief of steric hindrance which is the driving force in the silyl
migration.5 In the ring-opening reaction in the presence of
triethylamine, the triethylammonium cation presumably acts as
a proton source, thus neutralising the C-2 alkoxide and thereby
preventing silyl migration.
In order to explore this reaction further, we also examined the
reaction of the epoxide 1 with thiophenol itself (Scheme 2). The
major product from this reaction was the a-phenylthioglycoside
2a (36%), together with a minor amount of the b-phenylthiogly-
coside 2b (28%), already identified. Although the yields are not
very high, the process has not been optimised, especially with
respect to solvent. In this case, the reaction probably proceeds
through the intermediacy of an oxonium ion, formed by ring
opening of the epoxide, which is then selectively trapped from
the a-face.
The adduct 2a was crystalline, and we were able to confirm
its structure by X-ray diffraction, as indicated in Fig. 1.† We
were surprised to discover a 1C4 conformation in the crystalline
state, in which the four substituents at C-2, C-3, C-4 and C-5 are
in axial positions, with the phenylthio group equatorial. The
other notable feature is that there is a hydrogen bond between
the C-2 hydroxy proton and O-1, which may help to stabilise the
1C4 conformation. The 1C4 conformation does appear to be very
rare, although it has been observed in the case of a bis(tert-
butyldiphenylsilyl) a-aryl C-olivoside, in which three sub-
stituents, rather than four, are in an axial position.6 Analysis of
the 1H NMR spectrum of 2a suggested that, in CDCl3 at least,
the solution conformation was very similar to that in the solid
O(4)
Si(2)
O(2)
O
SPh
O
ButMe2SiO
ButMe2SiO
ButMe2SiO
ButMe2SiO
O(1)
C(2)
C(4)
i
O
S(1)
OSiMe2But
C(3)
C(5)
C(1)
C(6)
OR
OSiMe2But
3 R = H
1
O(3)
iii
4 R = SiMe2But
ii
O(5)
Si(3)
O
SPh
OH
ButMe2SiO
ButMe2SiO
Si(1)
OSiMe2But
2b
Scheme 1 Reagents and conditions: i, PhSLi (from PhSH and BuLi), 0 °C
to room temp., THF, 16 h; ii, PhSH, NEt3, 0 °C to room temp., 16 h, THF;
iii, ButMe2SiOSO2CF3, 2,6-lutidine, CH2Cl2, 278 °C to room temp., 16 h
Fig. 1 The conformation of 2a in the crystal structure. Hydrogen atoms,
other than the hydroxy proton, and minor disorder components are omitted
for clarity.
Chem. Commun., 1997
1855

View Article Online
displacement parameters, riding H atoms (with rotational freedom for
state, with very small vicinal coupling constants around the
ring, and significant long-range (W) coupling between H-2 and
H-4, and between H-3 and H-5 (indicating that each of these
protons was in an equatorial position). The corresponding
sulfone 5, prepared by oxidation of the sulfide 2a, exhibited
almost identical coupling constants and is therefore probably
also in the same conformation in solution.
methyl and hydroxy hydrogen atoms), and two-fold disorder for one silyl
1
2
2 2
group. Rw = [Sw(Fo 2 Fc2)/Sw(Fo ) ] = 0.223 for all data, conventional
2
R = 0.101 for F values of 5551 reflections with Fo > 2s(Fo2), goodness
2
of fit 1.129, absolute configuration indicated by refinement of enantiopole
parameter (ref. 9) to 20.3(2), residual electron density within ± 0.61 e Å23
.
The indicated absolute configuration was consistent with the known
absolute configuration of the starting material. The quality of data collection
and structure refinement are limited by the deformed nature of the crystals
(apparently slightly bent, giving distorted reflections) and the structural
disorder. Programs were standard Siemens control and integration software,
Siemens SHELXTL, and local programs. CCDC 182/580.
The ‘A value’ of the phenylthio group (4.60–5.19 kJ mol21 7
)
is not sufficiently large to force the other substituents axial, so
the most obvious explanation is that the adjacent tert-
butyldimethylsilyloxy groups at C-3 and C-4 are more sterically
encumbered when they are diequatorial than when they are
diaxial.‡ Eliel’s important report on the ‘A values’ of silyloxy
groups, which clearly showed (rather counter-intuitively) that a
tert-butyldimethylsilyloxy group had a smaller ‘A value’ than a
trimethylsilyloxy group,8 provides strong support for this idea.
The ease with which the tert-butyldimethylsilyl group at C-3 of
2b migrates to C-2 is probably another manifestation of this
same phenomenon.
It is clear that the pronounced conformational effects of tert-
butyldimethylsilyl ethers may have useful consequences in
control of reaction stereoselectivity.
We thank the EPSRC for a CASE award (C. L. W.) and for
equipment grants, Rhoˆne-Poulenc Rorer for support and Drs B.
Porter, A. J. Ratcliffe and M. Podmore (RPR) for helpful
discussions. R. F. W. J. also thanks the Nuffield Foundation for
a One Year Science Research Fellowship.
‡ This suggestion was first made by Suzuki (ref. 6). This type of behaviour
has been noted before in the case of glycals (ref. 10), and in b-lactams
derived from glycals by cycloaddition (ref. 11). It has also been suggested
as an influence on reactivity, but without direct evidence (ref. 4.).
1 R. L. Halcomb and S. J. Danishefsky, J. Am. Chem. Soc., 1989, 111,
6661.
2 C. H. Marzabadi and C. D. Spilling, J. Org. Chem., 1993, 58, 3761.
3 D. M. Gordon and S. J. Danishefsky, Carbohydr. Res., 1990, 206,
361.
4 For a related observation, see: T. Iimori, H. Takahashi and S. Ikegami,
Tetrahedron Lett., 1996, 37, 649.
5 A closely related migration of a tert-butyldimethylsilyl group has been
reported: R. W. Friesen and A. K. Daljeet, Tetrahedron Lett., 1990, 31,
6133. For an analogous migration of a trimethylsilyl group, see
V. Pedretti, A. Veyrie`res and P. Sina¨y, Tetrahedron, 1990, 46, 77.
6 T. Hosoya, Y. Ohashi, T. Matsumoto and K. Suzuki, Tetrahedron Lett.,
1996, 37, 663.
7 O. A. Subbotin, V. A. Palyulin, S. I. Kozhushkov and N. S. Zefirov,
J. Org. Chem. USSR (Engl. Transl.), 1978, 14, 196.
8 E. L. Eliel and H. Satici, J. Org. Chem., 1994, 59, 688.
9 H. D. Flack, Acta Crystallogr., Sect. A, 1983, 39, 876.
10 D. Horton, W. Priebe and O. Varela, Int. Carbohydr. Symp. XIIth,
Utrecht, July 1–7, 1984, Abstr. D8.35, p. 519.
11 M. Chmielewski, Z. Kaluza, H. Dodziuk, K. Suwinska, D. Rosenbaum,
H. Duddeck, P. D. Magnus and J. C. Huffmann, Carbohydr. Res., 1990,
203, 183.
Footnotes and References
* E-mail: r.f.w.jackson@newcastle.ac.uk
† Crystal data for 2a: C30H58O5SSi3, Mr = 615.1, orthorhombic, P212121,
a = 27.1719(17), b = 10.9754(7), c = 12.3479(8) Å, V = 3682.4(4) Å3,
Z = 4, Dc = 1.109 g cm 23, m = 0.22 mm21 (Mo-Ka, l = 0.71073 Å),
T = 160 K. Measurements were made on a Siemens SMART CCD area
detector diffractometer. Data were collected by narrow-frame w rotation.
19 575 reflections, 2q @ 50°, 6467 unique reflections, Rint = 0.091, semi-
empirical absorption correction, transmission 0.84–0.97. Structure solution
was by direct methods, refinement on F2 for all data, with anisotropic
Received in Liverpool, UK, 27th June 1997; 7/04534D
1856
Chem. Commun., 1997