Sulfur atom configuration of sulfinyl galactofuranosides determines different reactivities in glycosylation reactions

Article abstract of DOI:10.1016/S0040-4039(00)00887-X

Sulfinyl β-D-galactofuranosides 3(R,S) and 9(R,S) were used as new hexofuranosyl donors in glycosylation reactions. Activation by trifluoromethanesulfonic anhydride involved different reaction pathways depending on the stereochemistry at the sulfur atom. Experimental results underlined a more suitable reactivity of 3(R) and 9(R) versus 3(S) and 9(S), respectively, for the synthesis of β-D-galactofuranosides. (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0040-4039(00)00887-X

Tetrahedron Letters 41 (2000) 5515±5519
Sulfur atom con®guration of sul®nyl galactofuranosides
determines di€erent reactivities in glycosylation reactions
Vincent Ferrieres,a,* Jerome Joutel,a Rachel Boulch,a Myriam Roussel,a
Loõc Toupetb and Daniel Plusquelleca,*
a
Â
Á
Â
Ecole Nationale Superieure de Chimie de Rennes, Laboratoire de Syntheses et Activations de Biomolecules,
Â
 Â
Associe au CNRS, Avenue du General Leclerc, 35700 Rennes, France
b
Á
Â
Â
Â
Â
 Â
Groupe Matiere Condensee et Materiaux, Associe au CNRS, Universite de Rennes 1, Avenue du General Leclerc,
35042 Rennes, France
Received 5 May 2000; accepted 1 June 2000
Abstract
Sul®nyl b-d-galactofuranosides 3(R,S) and 9(R,S) were used as new hexofuranosyl donors in glycosylation
reactions. Activation by tri¯uoromethanesulfonic anhydride involved di€erent reaction pathways depend-
ing on the stereochemistry at the sulfur atom. Experimental results underlined a more suitable reactivity of
3(R) and 9(R) versus 3(S) and 9(S), respectively, for the synthesis of b-d-galactofuranosides. # 2000
Elsevier Science Ltd. All rights reserved.
Keywords: glycosidation; sulfoxides; sul®nyl glycosides; hexofuranosides; galactofuranosides.
Since the discovery of glycosyl donor properties of sul®nyl glycosides,1 the well known sulf-
oxide method was widely exploited even for the glycosylation of unreactive substrates1,2 and for
the direct synthesis of b-d-mannopyranosides.3 Mechanistic studies have clearly underlined the
formation of di€erent intermediates4 6 and, as a consequence of these fundamental approaches,
new methods were developed in order to improve both diastereoselectivity and yield by limiting
the formation of side products.7 Nevertheless, little consideration has been directed towards the
reactivity of each epimer8 of the sulfoxide donor in glycosylation reactions. It is acknowledged
that both isomers are similarly activated and so mixtures of R and S sul®nyl glycosides are gen-
erally used without previous separation.9 Our interest in sul®nyl glycosides arose from the
requirement of new hexofuranosyl donors for the synthesis of glycoconjugates containing such
unusual residues. The aldohexosides presenting the furanose con®guration are found in parasites
of the Trypanosomatid family but are not present in mammalian cells,10 and therefore represent
interesting targets for the design of new drugs. In this context, we herein describe ®rst evidences,
* Corresponding author. Fax: 33 (0)2 99 87 13 48; e-mail: ferriere@ensc-rennes.fr
0040-4039/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(00)00887-X

5516
corroborated by crystal structures, that sul®nyl galactofuranosides possessing di€erent absolute
con®guration at the sulfur atom, present di€erent reactivities as glycosyl donors.
Peracetylated phenyl thiogalactofuranoside 2, actual precursor of sulfoxides 3a,b (Scheme 1),
was prepared as described in previous papers from this laboratory.11,12 Mild sulfoxidation of
thiogalactofuranoside 2 with hydrogen peroxide and acetic acid in the presence of silica gel13 gave
a 1.7:1 mixture of both epimers in 71% yield. Structure of target 3a,b was con®rmed by NMR and
HRMS.14 Glycosylation of cyclohexanol (CyOH) as the model acceptor was then investigated
under Kahne conditions.1 Whilst, surprisingly, no reaction occurred at ^78ꢀC, the expected b-
galactofuranoside 4 was obtained after only 1 min at room temperature and was isolated in 63%
yield (Table 1, entry 1). Appreciable amounts of trehalose-like disaccharide 615 were also
obtained under slightly modi®ed conditions (entry 2) or from couplings using pure sulfoxide 3b
(entries 3±5) which could be chromatographically separated from its epimer 3a. In contrast,
orthoester 515 was mainly obtained from the major less polar sul®nyl furanoside 3a (entries 6, 7).
Consequently, our results showed that the glycosylation pathway using a sul®nyl galactofurano-
side is dictated predominantly by (i) the reaction conditions, and more importantly (ii) the
stereochemistry at the sulfur atom. Unfortunately, the latter could not be easily determined by
X-ray analysis since neither 3a nor 3b could be obtained in crystalline form.16
Scheme 1. (i) See Ref. 12c; (ii) H2O2, AcOH, SiO2: 3 (71%, 3a/3bˆ1.7:1); 9: (84%, 9a/9bˆ1.7:1); (iii) CyOH, DTBMP,
Tf2O: 4±6 (see Table 1); 10 (35%); (iv) 1. NaOMe, MeOH; 2-BzCl, Py (47%); (v) 1. BzCl, Py (85%); 2. TFAA, H2SO4,
CH2Cl2; EtSH, BF3OEt2 (75%)
Tetra-O-benzoyl sul®nyl galactofuranoside 7 was therefore prepared by standard protecting
group manipulations starting from pure 3a. Eventually, we were rewarded by the growth of single
crystals of 7, which sulfur atom con®guration was revealed to be S (Fig. 1).17
Suitable crystals for X-ray crystallographic analysis were again obtained for ethyl su®nyl furano-
side 9, which presented the same sulfur con®guration as 7(S)17 (Fig. 1). Indeed, 9 was prepared

5517
Table 1
Reaction conditions for glycosylation of cyclohexanol (CyOH) using sulfoxide 3
Figure 1. ORTEP plot of the crystal structures of compounds 7(S) and 9(S). Thermal ellipsoids are drawn at 50%
probability
from 1 by standard benzoylation followed by acetolysis using tri¯uoroacetic anhydride (TFAA),
glycosylation of ethanethiol and mild oxidation. HRMS analysis cleanly showed the desired
sulfoxidation18 and 1H NMR indicated a R/S ratio of 1.7:1, as for 3a,b. However, while 9(S) was
partially crystallized out of the crude mixture, 9(R) and 9(S) could not be separated by chroma-
tography. Nevertheless, treatment of 1 equiv. of 9(R,S) (R/S=1.7:1) with 2 equiv. of cyclohexanol
under the conditions described above a€orded the expected furanoside 10. The moderate 35%
yield was the result of recovering of the donor in 25% yield. It is noteworthy that the mixture of

5518
starting sulfoxides was signi®cantly enriched with 9(R), since the R/S ratio reached 6:1. Conse-
quently, a marked di€erence in reactivity between furanosyl sulfoxide R and its S epimer was
again observed in glycosylation reactions.
On the basis of previous studies,1 7 results obtained herein could be partly rationalized. Firstly,
isolation of orthoester 5 clearly indicates stabilization of a preformed anomeric cation trapped by
cyclohexanol. However, isolation of the trehalose-like difuranoside 6 requires in situ formation of
a glycosidic acceptor. Upon increasing temperature, the latter may be obtained by rearrangement
of starting sulfoxide into anomeric sulfenate4 which then could act either as another glycosyl
donor or as a new nucleophile able to react with 3, to a€ord compound 6. Complementary
mechanistic studies are currently under investigation in order to determine the discriminating
factors in the stereochemical activation of both R and S sulfoxides.
Tables of atomic coordinates, bond lengths and bond angles have been deposited within the
Cambridge Crystallographic Data Centre, 12 Union Road, Cambridge, CB2 1EZ, UK.
Acknowledgements
We wish to thank Pierre Guenot, Centre Regional de Mesures Physiques de l'Ouest (Rennes),
for recording HRMS spectra and for his precious help in interpreting results.
References
1. Kahne, D.; Walker, S.; Cheng, Y.; Van Engen, D. J. Am. Chem. Soc. 1989, 111, 6881±6882.
2. Thompson, C.; Ge, M.; Kahne, D. J. Am. Chem. Soc. 1999, 121, 1237±1244.
3. (a) Crich, D.; Sun, S. J. Org. Chem. 1996, 61, 4506±4507; (b) Crich, D.; Dai, Z. Tetrahedron 1999, 55, 1569±1580.
4. Gildersleeve, J.; Pascal Jr., R. A.; Kahne, D. J. Am. Chem. Soc. 1998, 120, 5961±5969.
5. Crich, D.; Sun, S. J. Am. Chem. Soc. 1997, 119, 11217±11223.
6. Crich, D.; Dai, Z.; Gastaldi, S. J. Org. Chem. 1999, 64, 5224±5229.
7. Gildersleeve, J.; Smith, A.; Sakurai, K.; Raghavan, S.; Kahne, D. J. Am. Chem. Soc. 1999, 121, 6176±6182.
8. Khiar, N.; Alonso, I.; Rodriguez, N.; Jimenez-Barbero, J.; Nieto, O.; Cano, F.; Foces-Foces, C.; Martin-Lomas,
M. Tetrahedron Lett. 1997, 38, 8267±8270.
9. Crich, D.; Mataka, J.; Sun, S.; Lam, K. C.; Rheingold, A. L.; Wink, D. J. J. Chem. Soc., Chem. Commun. 1998,
2763±2764.
10. de Lederkremer, R. M.; Colli, W. Glycobiology 1995, 5, 547±552.
11. Ferrieres, V.; Bertho, J. N.; Plusquellec, D. Carbohydr. Res. 1998, 311, 25±35.
12. (a) Gelin, M.; Ferrieres, V.; Plusquellec, D. Carbohydr. Lett. 1997, 2, 381±388; (b) Ferrieres, V.; Gelin, M.; Boulch,
R.; Toupet, L.; Plusquellec, D. Carbohydr. Res. 1998, 314, 79±83; (c) Gelin, M.; Ferrieres, V.; Plusquellec, D. Eur.
J. Org. Chem. 2000, 8, 1423±1431.
13. Kakarla, R.; Dulina, R. G.; Hatzenbuhler, N. T.; Hui, Y. W.; So®a, M. J. J. Org. Chem. 1996, 61, 8347±8349.
14. For [C20H24O10S+H]+: theoretical: 457.1168; experimental: 457.1149.
15. Structure of symmetrical difuranoside 6 and that of orthoester 5 were assigned on the basis of spectroscopic and
HRMS data.
16. During the course of this work, an NMR approach for predicting absolute con®guration at sulfur atom of
glycopyranosyl sulfoxides was published: Buist, P. H.; Behrouzian, B.; MacIsaac, K. D.; Cassel, S.; Rollin, P.;
Imberty, A.; Gautier, C.; Perez, S.; Genix, P. Tetrahedron: Asymmetry 1999, 10, 2881±2889.
17. Crystal data for 7(S): C40H32O10S, CH3OH, M=736.76, monoclinic, C2, a=28.357(5), b=11.120(4), c=12.509(3)
A, ꢀ=110.98(2)ꢀ, V=3683(2) A^3, Z=2, DX=1.329 Mg m^3. The data collection gives 4234 unique re¯ections
from which 2880 with I>2.0s(I). After Lorenz and polarization corrections the structure was solved with SIR-97,
which reveals the non-hydrogen atoms of the structure and a methanol molecule. After anisotropic re®nement,

5519
many hydrogen atoms are found with a Fourier Di€erence. The whole structure was re®ned with SHELXL97 by
the full-matrix least-square techniques {use of F magnitude; x, y, z, bij for S, C and O atoms, x, y, z in riding mode
for H atoms; 469 variables and 2880 observations with I>2.0s(I); calc. w=1/[s2(FO2 )+(0.0915P)2+0.588P] where
P=(F2O+2F2C)/3} with the resulting R=0.487, RW=0.128 and SW=1.033 (residual Ár<0.065 eA^3).
18. For [C36H32O10S+H]+: theoretical: 657.1794; experimental: 657.1799; Crystal data for 9(S): C36H32O10S, CH3OH,
M=656.68, orthorhombic, P212121, a=11.058(5), b=12.035(6), c=26.839(6) A, V=3572(3) A^3, Z=4, DX=1.221
Mg m^3. The data collection gives 4334 unique re¯ections from which 2203 with I>2.0s(I). After Lorenz and
polarization corrections the structure was solved with SIR-97 which reveals the non-hydrogen atoms of the
structure and a methanol molecule. After anisotropic re®nement, all the hydrogen atoms are found with a Fourier
Di€erence. The whole structure was re®ned with SHELXL97 by the full-matrix least-square techniques {use of F
magnitude; x, y, z, bij for S, C and O atoms, x, y, z in riding mode for H atoms; 425 variables and 4334
2
observations with I>2.0s(I); calc w ˆ 1=‰ꢁ2ꢁFꢁ2†‡ꢁ0:0728P† ‡0:0940PŠ where P=(FO2 +2F2C)/3} with the resulting
R=0.061, RW=0.1483 and SW=0.878 (residual Ár<0.32 eA^3).