Stereoselective synthesis of nucleosides from 1-thio and 1-seleno glycosides through consecutive 1,2-migration and glycosylation under Mitsunobu conditions

Article abstract of DOI:10.1016/S0040-4039(99)02071-7

2-Deoxy-2-sulfenyl-arabino-α-nucleosides and 2-deoxy-2-sulfenyl-ribo- β-nucleosides were obtained with excellent stereoselectivity from 1-thio- ribo- and 1-thio-arabino-glycosides under Mitsunobu conditions.

Full text of DOI:10.1016/S0040-4039(99)02071-7

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 41 (2000) 407–411
Stereoselective synthesis of nucleosides from 1-thio and 1-seleno
glycosides through consecutive 1,2-migration and glycosylation
under Mitsunobu conditions
Antonio Viso, Nicolai Poopeiko † and Sergio Castillón ∗
Departament de Química Analítica i Química Orgànica, Facultat de Química, Universitat Rovira i Virgili, Pça,
Imperial Tarraco 1, 43005 Tarragona, Spain
Received 22 September 1999; accepted 2 November 1999
Abstract
2-Deoxy-2-sulfenyl-arabino-α-nucleosides and 2-deoxy-2-sulfenyl-ribo-β-nucleosides were obtained with ex-
cellent stereoselectivity from 1-thio-ribo- and 1-thio-arabino-glycosides under Mitsunobu conditions. © 2000
Elsevier Science Ltd. All rights reserved.
A general problem in the synthesis of 20-deoxy- and 20,30-dideoxy-nucleosides is controlling the
stereoselectivity in the formation of the glycosidic bond. To overcome this lack of stereoselectivity, we
used sugars with easily removable electron donor groups at position 2 of the sugar ring as glycosyl
donors (e.g. phenylsulfenyl1 and phenylselenenyl2 to give the corresponding 2-phenylsulfenyl and 2-
phenylselenenyl nucleosides). Phenylsulfenyl and phenylselenenyl groups were introduced into the
sugar ring by reacting PhSCl or PhSeCl with the corresponding lactone enolate followed by reduc-
tion of the lactone (Scheme 1, via 1). An alternative procedure starts from glycals, and successive
NIS, I2, PhSCl or PhSeCl addition and glycosylation (Scheme 1, via 2, X=I, SPh, SePh, respec-
tively) leads to the corresponding 20,30-dideoxy-20-iodo-,3 20,30-dideoxy-20-phenylsulfenyl-,4 and 20,30-
dideoxy-20-phenylselenenyl nucleosides5 with good to excellent stereoselectivities. Recently, we obtain-
ed 2,30-dideoxy-30-fluoro-β-nucleosides by converting phenyl 3-deoxy-3-fluoro-1-seleno-α-arabino-
furanosides to 20,30-dideoxy-30-fluoro-20-phenylselenenyl-β-ribo nucleosides through consecutive 2-OH
activation, 1,2-migration and glycosylation in Mitsunobu conditions (Scheme 1, via 3) and subsequent
deselenization.6 An epi-ion (B) (Scheme 1) has been proposed in these cases although some reports doubt
this proposal, suggesting that oxonium cation A is the real intermediate.7
∗
†
Corresponding author.
Permanent address: Institute of Bioorganic Chemistry, National Academy of Sciences, Kuprevicha 5, 220141 Minsk,
Belarus.
0040-4039/00/$ - see front matter © 2000 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(99)02071-7
tetl 15999

408
Scheme 1.
In this report we show that 1-thio-β-ribo- and 1-thio-α-arabino-glycosides are appropriate starting
materials for preparing 20-deoxy-20-sulfenyl-arabino-α-nucleosides and 20-deoxy-20-sulfenyl-ribo-β-
nucleosides, respectively, which are precursors of 20-deoxy nucleosides, with excellent stereoselectivity.
The selenoglycoside 2 and thio-glycosides 3–6 were prepared by reacting tetra-O-acetyl-D-ribose (1)
with phenyl selenol or the corresponding thiols in the presence of BF3·OEt2 and subsequent deprotection
and protection with TIPDSCl (Scheme 2).
Scheme 2.
Initially we tested the reaction of the 1-seleno-ribo-glycoside 2 with 3-benzoylthymine (N3-BzThy)
under Mitsunobu conditions (PPh3/DEAD) in DMF as the solvent, and we obtained the arabino-
nucleoside 7 in a 60% yield and an α:β ratio of 3:1 (Table 1, entry 1). This means that after the OH
has been activated there is a 1,2-transposition and then the base enters at position 1 (see Scheme 1,
via 3).8 When the 1-thio-glycoside 3, a more useful starting material, reacted in similar conditions, the
reaction was very slow; after 72 h it still had not finished. However, when the reaction was carried out at
50°C, nucleoside 8 was obtained in a 65% yield and with low stereoselectivity in 30 min (entry 2). The
reaction was then tried in other solvents such as THF at low (entry 3) and high temperatures (entry 4). The
resulting conversions and selectivities were always poor. When dichloroethane was used as the solvent
the yield increased but the stereoselectvity was nil (entry 5). In all cases we obtained α/β mixtures of
nucleosides with an arabino configuration. This suggests that the reaction intermediate is not a thiiranium
cation but an oxonium cation.
To promote the formation of the thiiranium cation we prepared the 1-thioglycoside 4 with higher
electron density in the sulfur. However, when it was used as the starting material in the glycosylation
neither the yield nor the selectivity improved (entry 6). Then we synthesized the 1-thio-glycosides 5 and
6 with a bulkier substituent at the anomeric position. Compound 5 did not react even after 5 h at 50°C

409
Table 1
Synthesis of 2-phenylselenenyl- and 2-phenylsulfenyl-arabino-nucleosides from 1-seleno- and 1-thio-
ribo-glycosidesa
(entry 7). tert-Butyl 1-thio-glycoside 6, however, gave nucleoside 11 in a 64% yield and a ratio α:β=10:1
(entry 8). In this case results were best in DMF at room temperature. Thus, when the temperature was
increased (entry 9) or the solvent changed (entry 10), the yield and the stereoselectivities were worse.
In the previous reactions no O-glycosylation product was obtained, as has been observed in other
Mitsunobu reactions involving pyrimidinic bases.9
Taking into account the results obtained with the 1-seleno and 1-thio-ribo-glycosides, we prepared the
1-seleno-arabino-glycoside 13 and the 1-thio-arabino-glycosides 14 and 15, which have the anomeric
substituents identical to the ones that the best results in the ribo series (e.g. compounds 2–6) (Scheme 3).
In these cases the 1,2-transposition of the anomeric substituents should take place on the more hindered
α face to provide nucleosides with a ribo configuration.
Scheme 3.
The reaction of the 1-seleno-arabino-glycoside 13 with N3-BzThy in DMF at room temperature gave
nucleoside 16 in a 48% yield and an α:β ratio of 1:7 (Table 2, entry 1). As expected, only nucleosides with
ribo configuration were obtained, the major one being the β anomer. When the phenyl 1-thio-glycoside
14 was used as starting material the yield of nucleoside 17 improved to 71% and the α:β ratio increased
to 1:10 (entry 2). Interestingly, starting from the tert-butyl 1-thio-glycoside 15, anomer β was obtained
as the only product in a 52% yield, although in this case the reaction was slower (entry 3). Solvents
other than DMF (such as THF) required more drastic conditions and provided very low yields, although
the stereoselectivity was also excellent (entry 4). Starting from 15 and using 6-chloro-purine as base,
nucleoside 19 was obtained with excellent stereoselectivity. DMF probably plays two different roles in

410
this process; it solubilizes the pyrimidinic base and stabilizes the oxonium intermediate, thus increasing
the reaction rate.
Table 2
Synthesis of 2-phenylselenenyl- and 2-phenylsulphenyl-ribo-nucleosides from 1-seleno- and 1-thio-
arabino-glycosidesa
The structure of nucleosides was determined by NMR spectroscopy. Characteristic signals at 49–60
ppm in 13C NMR are indicative of the presence of RSe and RS groups at C-2. The configuration of the
nucleosides obtained could not be determined by the coupling constants values. NOE experiments shown
below confirm the proposed stereochemistry.
In conclusion, 2-deoxy-2-sulfenyl-arabino-α-nucleosides and 2-deoxy-2-sulfenyl-ribo-β-nucleosides
were obtained with excellent stereoselectivity from 1-thio-ribo- and 1-thio-arabino-glycosides in a one
pot reaction under Mitsunobu conditions. The stereoselectivity depends on the substituent bonded to the
sulfur and on the reaction conditions, particularly the solvent. From the results obtained it is apparent
that the reaction intermediate is an oxonium cation and not the thiiranium cation.
Acknowledgements
Financial support by DGICYT PB95-521-A (Ministerio de Educacion y Cultura, Spain) is acknowled-
ged. Technical assistance by the Servei de Recursos Cientifics (URV) is acknowledged.
References
1. (a) Wilson, L. J.; Liotta, D. C. Tetrahedron Lett. 1990, 31, 1815. (b) Kawakami, K.; Ebata, T.; Koseki, K.; Matsushita, H.;
Naoi, Y.; Itoh, K. Chem. Lett. 1990, 1459.
2. (a) Warren Beach, J. W.; Kim, H. O.; Jeong, L. S.; Nampalli, S.; Islam, Q.; Ahn, S. K.; Babu, J. R.; Chu, C. K. J. Org. Chem.
1992, 57, 3887. (b) Chu, C. K.; Babu, J. R.; Beach, J. W.; Ahn, S. K.; Huang, H.; Jeong, L. S.; Lee, S. J. J. Org. Chem. 1990,
55, 1418.
3. (a) McDonald, F. E.; Gleason, M. M. Angew. Chem., Int. Ed. Engl. 1995, 34, 350. (b) Kim, C. U.; Misco, P. F. Tetrahedron
Lett. 1992, 33, 5733.
4. (a) Wang, J.; Wurster, J. A.; Wilson, L. J.; Liotta, D. Tetrahedron Lett. 1993, 34, 4881. (b) Kawakami, K.; Ebata, T.; Koseki,
K.; Matsushita, H.; Naoi, Y.; Itoh, K. Heterocycles 1993, 36, 665. (c) Kawakami, K.; Ebata, T.; Koseki, K.; Matsushita, H.;
Naoi, Y.; Itoh, K. Heterocycles 1993, 36, 2765.

411
5. (a) El-Laghdach, A.; Díaz, Y.; Castillón, S. Tetrahedron Lett. 1993, 34, 2821. (b) El-Lagdach, A.; Matheu, M. I.; Castillón,
S. Tetrahedron 1994, 50, 12219. (c) Díaz, Y.; El-Laghdach, A.; Matheu, M. I.; Castillón, S. J. Org. Chem. 1997, 62, 1501.
6. Poopeiko, N.; Fernández, R.; Barrena, M. I.; Castillón, S.; Forniés-Cámer, J.; Cardin, C. J. J. Org. Chem. 1999, 64, 1375.
7. (a) Jones, D. K.; Liotta, D. C. Tetrahedron Lett. 1993, 34, 7209. (b) Dudley, T. J.; Smoliakova, I. P.; Hoffmann, M. R. J. Org.
Chem. 1999, 64, 1247. (c) Fachini, M.; Lucchini, V.; Modena, G.; Pasi, M.; Pasquato, L. J. Am. Chem. Soc. 1999, 121, 3944.
8. Rearrangements of 1-thio-glycosides in glycoside synthesis have been described: (a) Nicolaou, K. C.; Ladduwerhetty, T.;
Randall, J. L.; Chucholowsky, A. J. Am. Chem. Soc. 1986, 108, 2466. (b) Zuurmond, H. M.; van der Klein, P. A. M.; van der
Marel, G. A.; van Boom, J. H. 1993, 49, 6501.
9. Díaz, Y.; Bravo, F.; Castillón, S. J. Org. Chem. 1999, 64, 6508.
10. Typical experimental procedure: The base (0.38 mmol), Ph3P (0.3 mmol), DEAD (0.3 mmol) in DMF (2.5 ml) were added
under vigorous stirring to a mixture of the thio- or seleno-glycoside (0.2 mmol) which had been pre-dried by heating at 40°C
under vacuum. After the reaction was complete, the reaction mixture was diluted with EtOAc (25 ml) and washed with H2O
(3×10 ml). The combined H2O extracts were washed with EtOAc (10 ml) and the organic extracts were combined, dried,
and evaporated to dryness. The residue was chromatographed using MPLC with a linear EtOAc gradient in hexane (0–75%
v/v).