3-Deoxy-3,3-difluoro-D-arabinofuranose: First Stereoselective Synthesis and Application in Preparation of gem-Difluorinated Sugar Nucleosides

Article abstract of DOI:10.1021/jo034512i

The design and synthesis of gem-difluorinated sugar nucleosides were described. The key intermediate, 3-deoxy-3,3-difluoro-D-arabinofuranose 9, was first stereoselectively prepared from the chiral gem-difluorohomoallyl alcohol 12. The kinetic formation of

Full text of DOI:10.1021/jo034512i

3-Deoxy-3,3-d iflu or o-D-a r a bin ofu r a n ose: F ir st Ster eoselective
Syn th esis a n d Ap p lica tion in P r ep a r a tion of gem -Diflu or in a ted
Su ga r Nu cleosid es
Xingang Zhang,† Hairong Xia,† XiCheng Dong,† J ing J in,† Wei-Dong Meng,‡ and
Feng-Ling Qing*,†,‡
Key Laboratory of Organofluorine Chemistry, Shanghai Institute of Organic Chemistry, Chinese Academy
of Science, 354 Fenglin Lu, Shanghai 200032, China, and College of Chemistry and Chemistry
Engineering, Donghua University, 1882 West Yanan Lu, Shanghai 200051, China
flq@pub.sioc.ac.cn
Received April 22, 2003
The design and synthesis of gem-difluorinated sugar nucleosides were described. The key
intermediate, 3-deoxy-3,3-difluoro-D-arabinofuranose 9, was first stereoselectively prepared from
the chiral gem-difluorohomoallyl alcohol 12. The kinetic formation of single anti-14 in the benzylation
of 12 could be accomplished by controlling the amount of sodium hydride used. The dihydroxylation
of 14 (a mixture of anti and syn isomers) followed by deprotection and oxidation stereoselectively
afforded furanose 9 with the arabino configuration at the C2 position. N1-(3-Deoxy-3,3-difluoro-â-
D-arabinofuranosyl)cytosine 6 was prepared from 9 by the glycosylation reaction. 4′-Thiofuranose
25 was easily synthesized from 9. The oxidation of 25 followed by the condensation with silylated
N4-benzoylcytosine (Pummerer reaction) failed to give our desired protected nucleoside L-3′-deoxy-
3′,3′-difluoro- 4′-thiocytidine 27′, but the regioisomer 27 was obtained. The regiochemistry of the
Pummerer reaction was determined by the kinetic acidity of the R-proton of 4′-thiofuranose 25.
In tr od u ction
isopolar and isosteric substituent for oxygen.4 Analogues
of di- and triphosphates in which the CF2 group has
replaced the pyrophosphate oxygen have been used as
substrates in enzymatic reactions.4c,d,5 Since then, the CF2
group was used extensively to modify not only nucleotide
but also nucleoside analogues. For example, 2′-deoxy-
2′,2′-difluorocytidine (gemcitabine, 3, Figure 1) has been
approved as a drug for solid tumor treatment.6 However,
the 3′-deoxy-3′,3′-difluoro nucleosides have been much
less studied,7 which is probably due to the shortcomings
of existing synthetic methods. Recently, a number of the
L-configuration nucleoside analogues, the enantiomers of
the natural D-nucleosides, have emerged as potent anti-
The demand for new antiviral and anticancer agents
has led to the discovery of a class of modified nucleosides.
Fluorinated nucleosides are attractive compounds, with
the fluorine incorporated either into the nucleobase or
sugar moiety. The introduction of fluorine into the sugar
moiety of some nucleosides resulted in compounds with
a broad spectrum of antiviral and anticancer activities.1
Fluorine atom is a good mimic of a proton (small size) or
hydroxyl group (similar polarity). Many nucleoside ana-
logues in which fluorine is substituted for hydrogen or
hydroxyl on a carbohydrate carbon are known and of
considerable interest because of their biological activities.
For example, replacement of the 2′-â-hydrogen atom
(arabino configuration) or the 3′-hydroxyl group of natu-
ral thymidine by fluorine afforded new nucleosides with
potent antiviral properties, FMAU2 (1, Figure 1) and
FLT3 (2, Figure 1), respectively. The gem-difluorometh-
ylene (CF2) has been suggested by Blackburn as an
(3) (a) Etzold, G.; Hintsche, R.; Kowollik, G.; Langen, P. Tetrahedron
1971, 27, 2463. (b) Fleet, G. W. J .; Son, J . C.; Derome, A. E.
Tetrahedron 1988, 44, 625. (c) Balzarini, J .; Baba, M.; Pauwels, R.;
Herdewijin, P.; De Clercq, E. Biochem. Pharmacol. 1988, 37, 2847.
(4) (a) Blackburn, G. M.; England, D. A.; Kolkmann, F. J . Chem.
Soc., Chem. Commun. 1981, 930. (b) Blackburn, G. M.; Brown, D.;
Martin, S. J . J . Chem. Res., Synop. 1985, 92. (c) Blackburn, G. M.;
Eckstein, F.; Kent, D. E.; Perree, T. D. Nucleosides Nucleotides 1985,
4, 165. (d) Blackburn, G. M.; Kent, D. E. J . Chem. Soc., Perkin Trans.
1 1986, 913. (e) Blackburn, G. M.; Peree, T. D.; Rashid, A.; Bisbal, C.;
Lebleu, B. Chem. Scrip. 1986, 26, 21. (f) Blackburn, G. M.; Brown, D.;
Martin, S. J .; Parratt, M. J . J . Chem. Soc., Perkin Trans. 1 1987, 181.
(5) (a) Plantier-Royon, R.; Portella, C. Carbohydr. Res. 2000, 327,
119. (b) Tozer, M. J .; Herpin, T. F. Tetrahedron 1996, 52, 8619. For
examples, see: (c) Thatcher, G. R. J .; Campbell, A. S. J . Org. Chem.
1993, 58, 2271. (d) Chambers, R. D.; J aouhari, R.; O’Hagan, D.
Tetrahedron 1989, 45, 2272. (e) Halazy, S.; Ehrhard, A.; Danzin, C. J .
Am. Chem. Soc. 1991, 113, 315.
* To whom correspondence should be addressed. Fax: 86-21-
64166128.
† Shanghai Institute of Organic Chemistry.
‡ Donghua University.
(1) (a) Ternansky, R. J .; Hertel, L. W. In Organofluorine Compounds
in Medicinal Chemistry and Biomedical Applications; Filler, R.,
Kobayashi, Y., Yagupolski, L. M., Eds.; Elsevier: Amsterdam, 1993; p
23. (b) Herdewijn, P.; Van Aerschot, A.; Kerremans, L. Nucleosides
Nucleotides 1989, 8, 65. (c) Tsuchiya, T. Adv. Carbohydr., Chem.
Biochem. 1990, 48, 91. (d) Pankiewicz, K. W. Carbohydr. Res. 2000,
327, 87.
(6) (a) Hertel, L. W.; Kroin, J . S.; Misner, J . W.; Tustin, J . M. J .
Org. Chem. 1988, 53, 2406. (b) Hertel, L. W.; Boder, G. B.; Kroin, J .
S.; Rinzel, S. M.; Poore, G. A.; Todd, G. C.; Grindey, G. B. Cancer Res.
1990, 50, 4417.
(2) Watanabe, K. A.; Reichman, U.; Hirota, K.; Lopez, C.; Fox, J . J .
J . Med. Chem. 1979, 22, 21.
10.1021/jo034512i CCC: $25.00 © 2003 American Chemical Society
Published on Web 10/22/2003
9026
J . Org. Chem. 2003, 68, 9026-9033

3-Deoxy-3,3-difluoro-D-arabinofuranose
F IGURE 1. Rationale for the design of the target molecules 6 and 7.
viral agents against HIV and HBV.8 L-Nucleosides are
generally endowed with a lower host toxicity while
maintaining good antiviral activity in comparison with
their corresponding D-nucleosides. One of the first L-
nucleoside derivatives, L-2′,3′-dideoxy-3′-thiacytidine (lami-
vudine, 4; Figure 1), has been approved by FDA for use
in combination therapy against HIV and HBV.9 4′-
Thionucleosides, in which the furanose ring oxygen is
replaced by a sulfur atom, have been studied extensively
over the past 10 years because of their potent biological
activity.10 Notably, the resistance of the glycosidic bond
of 4′-thionucleosides to enzymatic hydrolysis catalyzed
by nucleoside phosphorylase is one of the advantages as
nucleoside antiviral agents.11 This has been applied to
SCHEME 1
(7) To our knowledge, only 2′,3′-deoxy-3′,3′-difluoro nucleosides were
prepared by the nucleophilic fluorination the corresponding precursors
with (diethylamino)sulfur trifluoride (DAST), but the yield was low.
(a) Bergstrom, D.; Romo, E.; Shum, P. Nucleosides Nucleotides 1987,
6, 53. (b) Bergstrom, D. E.; Mott, A. W.; De Clercq, E.; Balzarini, J .;
Swartling, D. J . J . Med. Chem. 1992, 35, 3369. (c) Gumian, G.;
Schinazi, R. F.; Chu, C. K. Org. Lett. 2001, 3, 4177.
(8) (a) Nair, V.; J ahnke, T. S. Antimicrob. Agents Chemother. 1995,
39, 1017. (b) Wang, P.; Hong, J . H.; Cooperwood, J . S.; Chu, C. K.
Antiviral Res. 1998, 40, 19.
(9) (a) Schinazi, R. F.; Chu, C. K.; Peck, A.; McMillan, A.; Mathis,
R.; Cannon, D.; J eong, L.-S.; Beach, J . W.; Choi, W.-B.; Yeola, S.; Liotta,
D. C. Antimicrob. Agents Chemother. 1992, 36, 672. (b) Coates, J . A.
V.; Cammack, N.; J enkinson, H. J .; J owett, A. J .; J owett, M. I.; Pearson,
B. A.; Penn, C. R.; Rouse, P. L.; Viner, K. C.; Cameron, J . M.
Antimicrob. Agents Chemother. 1992, 36, 733. (c) Doong, S.-L.; Tsai,
C.-H.; Schinazi, R. F.; Liotta, D. C.; Chen, Y.-C. Proc. Natl. Acad. Sci.
U.S.A. 1991, 88, 8495. (d) Beach, J . W.; J eong, L. S.; Alves, A. J .; Pohl,
D.; Kim, H. O.; Chang, C.-N.; Doong, S.-L.; Schinazi, R. F.; Cheng,
Y.-C.; Chu, C. K. J . Org. Chem. 1992, 57, 2217. (e) Dienstag, J . L.;
Perrillo, R. P.; Schiff, E. R.; Bartholomev, M.; Vicary, C.; Rubin, M.
New Engl. J . Med. 1995, 333, 1657.
(10) For examples, see: (a) Dyson, M. R.; Coe. P. L.; Walker, R. T.
J . Med. Chem. 1991, 34, 2782. (b) Secrist, J . A., III; Tiwari, K. N.;
Riordan, J . M.; Montgomery, J . A. J . Med. Chem. 1991, 34, 2361. (c)
Van Draanen, N. A.; Freeman, G. A.; Harvey, R.; Short, S. A.; J ansen,
R.; Szczech, G.; Koszalka, G. W. J . Med. Chem. 1996, 39, 538. (d)
Rahim, S. G.; Trivedi, N.; Bogunovic-Batchelor M. V.; Hardy, G. W.;
Mills, G.; Selway, J . W. T.; Snowden, W.; Littler, E.; Coe, P. L.; Basnak,
I.; Whale, R. F.; Walker, R. T. J . Med. Chem. 1996, 39, 789. (e)
Yoshimura, Y.; Kitano, K.; Yamada, K.; Satoh, H.; Watanabe, M.;
Miura, S.; Sakata, S.; Sasaki, T.; Matsuda, A. J . Org. Chem. 1997, 62,
3140. (f) Yoshimura, Y.; Watanabe, M.; Satoh, H.; Ashida, N.; Ijichi,
K.; Sakata, S.; Machida, H.; Matsuda, A. J . Med. Chem. 1997, 40, 2177.
(g) Secrist, J . A., III; Tiwari, K. N.; Shortnacy-Fowler, A. T.; Messini,
L.; Riordan, J . M.; Montgomery, J . A.; Meyers, S. C.; Ealick, S. E. J .
Med. Chem. 1998, 41, 3865.
the 2′-deoxy-2′,2′-difluorinated nucleosides; e.g., 2′-deoxy-
2′,2′-difluoro-4′-thiocytidine (5, Figure 1) has potent
activity against human T-cell leukemia CCRF-HSB-2
cells.10e,12
Based on the above consideration and our ongoing
efforts to develop new antiviral and anticancer agents,
we designed our target molecules N1-(3-deoxy-3,3-dif-
luoro-D-arabinofuranosyl)cytosine (6, Figure 1) and L-3′-
deoxy-3′,3′-difluoro- 4′-thiocytidine (7, Figure 1). The gem-
difluoromethylene (CF2) group was at the C3′ position of
compounds 6 and 7. Moreover, compounds 6 and 7 would
combine the characteristics of compounds 1-5 based on
the bioisosteric rationale (Figure 1). Our retrosynthetic
analysis (Scheme 1) was based on the thought that the
target molecule 7 could be derived from the precursor of
type 8, which could be used to introduce a base moiety
at the C1 position via Pummerer rearrangement.10e,13 We
envisioned that the gem-difluorinated furanose 9 would
be a suitable precursor for 8. At the same time, the target
compound 6 could be prepared from 9 by the glycosylation
(11) Parks, R. E., J r.; Stoeckler, J . D.; Cambor, C.; Savarese, T. M.;
Crabtree, G. W.; Chu, S.-H. In Molecular Actions and Targets for
Cancer Chemothreapeutic Agents; Sartorelli, A. C., Lazo, J . S., Bertino,
J . R., Eds.; Academic Press: New York, 1981; p 229.
(12) Yoshimura, Y.; Kitano, K.; Satoh, H.; Watanabe, M.; Miura,
S.; Sakata, S.; Sasaki, T.; Matsuda, A. J . Org. Chem. 1996, 61, 822.
J . Org. Chem, Vol. 68, No. 23, 2003 9027

Zhang et al.
SCHEME 2
SCHEME 3
reactions. The preparation of 9, however, appeared to be
a great challenge. The gem-difluoromethylene group is
frequently prepared through the difluorination of a
carbonyl group. The reagents commonly utilized for this
transformation are SF4,14 SeF4,15 MoF6,16 PhSF3,17 and
(diethylamino)sulfur trifluoride (DAST).18 Only DAST is
commercially available. However, very few sterically
hindered five-membered cyclic ketones have been diflu-
orinated by DAST. gem-Difluorinated furanose 9 has not
been synthesized via this method. Very recently, Chu et
al. reported difluororibose derivative 11 could not be
obtained from ketone 10 in the classic fluorinating
conditions (Scheme 1).7c Even when the reaction was run
in neat DAST, no products could be detected, and
attempts to increase the reactivity by raising the tem-
perature only caused decomposition of 10. Furthermore,
in the only example of difluorination of an R,R′-trans-
disubstituted five-membered cyclic ketone found in the
literature, the carbonyl group is relatively unhindered,
and the yield is low (25%).19 Moreover, the synthetic
method reported for this type is rare. As gem-difluori-
nated furanose 9 is a key intermediate in the synthesis
of our designed target molecules 6 and 7 (Scheme 1), it
is necessary to develop a new methodology for its
preparation. Herein, we describe the first efficient and
stereoselective route to 3-deoxy-3,3-difluoro-D-arabino-
furanose 9 and its application in the synthesis of gem-
difluorinated sugar nucleosides.
difluorohomoallyl alcohol 12 by the reaction of aldehyde
13 with (R,R-difluoroallyl)silane21 in the presence of
fluoride anion failed to produce the desired compound 12
(Scheme 3). Fortunately, using the gem-difluoroalllylin-
dium generated from 3-bromo-3,3-difluoropropene and
indium in DMF instead of (R,R-difluoroallyl)silane af-
forded compound 12 in high yield.22 The ratio of anti/
syn of compound 12 is 7.7/1 determined by 19F NMR. The
signal of the difluoromethylene group in anti-12 appeared
at higher field than that in syn-12 in the 19F NMR
spectrum.23 Notably, the anti-12 isomer is our desired
compound.
Benzylation of 12 was easily accomplished with sodium
hydride (1.6 equiv) and catalytic tetrabutylammonium
iodide, followed by benzyl bromide in 93% yield, but the
ratio of anti/syn of 14 was changed to 5.7/1 (Scheme 4).
This was due to the electron-withdrawing effect of the
difluoromethylene group, which could make the hydrogen
of the C3 position possess some acidity. The configuration
of the C3 position was slightly racemized under basic
conditions (see the following section for detailed discus-
sion). The Os-catalyzed dihydroxylation of compounds 14
gave the mixture of compound 15 and 16 in 95% yield
(15/16 ≈ 1/1). Compounds 15 and 16 could be separated
by column chromatography. Fortunately, when compound
14 was subjected to the reaction conditions of catalytic
asymmetric dihydroxylation (AD) of 3,3,3-trifluoropro-
pene reported by Sharpless et al. using (DHQ)2PYR as a
Resu lts a n d Discu ssion
Syn th esis of 3-Deoxy-3,3-d iflu or o-D-a r a bin ofu r a -
n ose 9. Scheme 2 illustrated that key intermediate 9 was
formally derived from the chiral gem-difluorohomoallyl
alcohol 12 through dihydroxylation and followed by ring
closure. To control the absolute configuration of the target
compound, one could perform the synthesis in an enan-
tioselective fashion. To obtain the gem-difluorohomoallyl
alcohol 12, the coupling of the gem-difluoroallylic metal
species with carbonyl compounds was investigated.20
1-(R)-Glyceraldehyde acetonide 13 was used as the
starting material. Initial attempts to synthesize the
(13) (a) O’Neil, I. A.; Hamilton, K. M. Synlett 1992, 791. (b)
Nishizono, N.; Yamagata, Y.; Matsuda, A. Tetrahedron Lett. 1996, 37,
7569. (c) Naka, T.; Nishizono, N.; Minakawa, N.; Matsuda, A. Tetra-
hedron Lett. 1999, 40, 6297. (d) Naka T.; Minakawa, N.; Abe, H.; Kaga,
D.; Matsuda, A. J . Am. Chem. Soc. 2000, 122, 7233.
(14) (a) Boswell, G. A., J r.; Ripka, W. C.; Scribner, R. M.; Tullock,
C. W.Org. React. (N. Y.) 1974, 21, 20-30. (b) Wang, C.-L. J . Org. React.
(N.Y.) 1985, 34, 319.
(15) Kent, P. W.; Wood, K. R. British Patent 1 136075, 1968.
(16) Mathey, F.; Bensoam, J . Tetrahedron 1971, 27, 3965.
(17) Sheppard, W. A. J . Am. Chem. Soc. 1962, 84, 3058.
(18) (a) Middleton, W. J . J . Org. Chem. 1975, 40, 574. (b) Hudlicky,
M. Org. React. (N.Y.) 1988, 35, 513.
(19) (a) Biggadike, K.; Borthwick, A. D.; Evans, D.; Exall, A. M.;
Kirk, B. E.; Roberts, S. M.; Stephenson, L.; Youds, P.; Slawin, A. M.
Z.; Williams, D. J . J . Chem. Soc., Chem. Commun. 1987, 251. (b)
Borthwick, A. D.; Evans, D. N.; Kirk, B. E.; Biggadike, K.; Exall, A.
M.; Youds, P.; Roberts, S. M.; Knight, D. J .; Coates, J . A. V. J . Med.
Chem. 1990, 33, 179
(20) For examples, see: (a) Seyferth, D.; Wursthorn, K. R. J .
Organomet. Chem. 1979, 182, 455. (b) Seyferth, D.; Simon, R. M.;
Sepelak, D. J .; Klein, H. A. J . Org. Chem. 1980, 45, 2273. (c) Seyferth,
D.; Simon, R. M.; Sepelak, D. J .; Klein, H. A. J . Am. Chem. Soc. 1983,
105, 4634. (d) Yang, Z.; Burton, D. J . J . Fluorine Chem 1989, 44, 339.
(e) Yang, Z.; Burton, D. J . J . Org. Chem. 1991, 56, 1037.
(21) (a) Hiyama, T.; Obayahsi, M.; Sawahata, M. Tetrahedron Lett.
1983, 24, 4113. (b) Fujita, M.; Hiyama, T. J . Am. Chem. Soc. 1985,
107, 4085. (c) Fujita, M.; Obayashi, M.; Hiyama, T. Tetrahedron 1988,
44, 4135.
(22) (a) Kirihara, M.; Takuwa, T.; Takizawa, S.; Momose, T.
Tetrahedron Lett. 1997, 38, 2853. (b) Kirihara, M.; Takuwa, T.;
Takizawa, S.; Momose, T.; Nemoto, H. Tetrahedron 2000, 56, 8275.
(23) Wirth, D. D. EP 0 727 433 A1 1996.
9028 J . Org. Chem., Vol. 68, No. 23, 2003

3-Deoxy-3,3-difluoro-D-arabinofuranose
SCHEME 4 . Syn th esis of 3-Deoxy-3,3-d iflu or o-D-a r a bin ofu r a n ose 9a
a
Key: (a) NaH (1.6 equiv), BnBr, TBAI, THF, 0 °C, 12 h; (b) (DHQ)2PYR, K2OsO2(OH)4, K3Fe(CN)6, K2CO3, t-BuOH/H2O (1:1), 24 h,
63% de, or OsO4, NMNO, acetone/H2O, rt, 2 d; (c) BzCl, Py, CH2Cl2, -78 °C, 2 h; (d) (i) 75% AcOH, 50 °C, 1.5 h, (ii) NaIO4, acetone/H2O,
rt., 1.5 h.
F IGUR E 2. ORTEP drawing of the X-ray crystallographic
structure of 6.
ligand,24 compound 15 was isolated in 60% yield (dr 15/
16 ) 4.4/1) (Scheme 4). Selective benzoylation at the
primary hydroxyl group of the resulting diol 15 at -78
°C gave benzoate 17 in 90% yield. The conversion of 17
to the desired intermediate 9 was achieved by the
F IGUR E 3. ORTEP drawing of the X-ray crystallographic
following two steps: (1) the acidic hydrolysis of the
isopropylidene group by treatment with 75% acetic acid
at 50 °C and (2) the oxidative scission of the resulting
diol with sodium periodate and subsequent cyclization.
Compound 9 was obtained in 94% yield from 17. It was
noteworthy that the compound 9 (>35/1, C2 (-OBn) dr
determined by 19F NMR before column) was formed in
structure of 8 and NOE correlations from NOESY spectra of
29a and 29b.
benzylation reaction conditions, the ratio of anti/syn was
changed from 7.7/1 (12) to 5.7/1 (14). To study the
epimerization of compound 12, different equivalents of
sodium hydride were used in the benzylation (Table 1).
Surprisingly, when the gem-difluorohomoallyl alcohol 12
reacted with less than 1.0 equiv of sodium hydride, only
one isomer anti-14 was obtained (entries 1 and 2).
Furthermore, the more sodium hydride was used, the less
compound 12 was recovered, and the more syn-12 was
enriched in the recovered 12. Notably, based on the yield
and the ratio of anti/syn-recovered 12 and formed 14,
respectively (entries 1-2), no epimerization happened in
recovered 12 and product 14. These results suggested
that this was a kinetic resolution and the reactivity of
anti-12 seemed to be much more active than that of syn-
12. So we can obtain the single anti-14 by controlling the
amount of sodium hydride used in the benzylation of
compound 12. The ratio of anti/syn-14 was similar to that
of anti/syn-12 in the case of 1.0 equiv of sodium hydride
1
two isomers. Both H NMR and 19F NMR spectra showed
the ratio of these two isomers was 1:1. This was a
stereospecific conversion resulting in the formation of the
furanose 9 withan arabino configuration at C2 position
(see the following section for detailed discussion). The C2
absolute configuration of 9 was determined by the X-ray
crystal structure of 6 and 8 (Figures 2 and 3). Thus, we
first succeeded in stereoselectively synthesizing the
3-deoxy-3,3-difluoro-D-arabinofuranose 9 from aldehyde
13 in an efficient way (34% overall yield, six steps).
Ep im er iza tion of Com p ou n d 12 in Ben zyla tion .
As shown in Scheme 4, when 12 was subjected to
(24) Vanhessche K. P. M.; Sharpless, K. B. Chem. Eur. J . 1997, 3,
517.
J . Org. Chem, Vol. 68, No. 23, 2003 9029

Zhang et al.
SCHEME 5. Rin g Op en in g a n d Acetyla tion of
TABLE 1. Ben zyla tion of 12 w ith Differ en t Equ iva len ts
of Sod iu m Hyd r id e
F u r a n ose 9
product 14 recovered 12
entry NaH (equiv)
anti/syna
anti/syna
yieldb (%)
1
2
3
4
5
0.6
0.8
1.0
1.2
1.6
21.0/1
21.8/1
8.2/1
6.1/1
5.7/1
2.8/1
1.8/1
(35.6);c 75.5
(25.0);c 78.5
80.0
87.4d
intermediates I and II formed from compound 17 were
in tautomeric equilibrium, and anomerization at this
carbon could occur easily due to the electron-withdrawing
effect of difluoromethylene (Scheme 6). The stereospecific
formation of the furanose 9 from intermediate I with the
arabino configuration at C2 position was thought to be
faster than the formation of furanose 9′ from intermedi-
ate II. At the same time, the intermediate II could be
easily transformed into intermediate I via conformer III
due to the energy deference between these two interme-
diates. In addition, the attack of the free hydroxyl group
of I on the carbonyl group (possible from both the re and
si face of the aldehyde moiety) resulted in a 1:1 mixture
of R/â anomers of 9.
93.1d
a
Determined by 19F NMR before column chromatography
purification. Isolated yield based on the NaH. c Recovery of the
substrate 12. Isolated yield based on the compound 12.
b
d
(entry 3). When 1.2 equiv of sodium hydride was used,
the epimerization happened, and the ratio of anti/syn-
14 was 6.1/1 (entry 4). Interestingly, when 1.6 equiv of
sodium hydride was used, further epimerization hap-
pened, but product 14 was isolated in the highest yield
(entry 5). No doubt, the epimerization of product 14 at
the C3 position did not happen until it was formed. The
more excess base was used, the more epimeric product
syn-14 was obtained. Although the gem-difluoromethyl-
ene (CF2) has been suggested as an isopolar and isosteric
substituent for oxygen, it is still an electron-withdrawing
group and thus could make the hydrogen of C3 position
in 14 with some acidity. The excessive base could make
the racemization happen.
Furthermore, the formation of 9 from intermediate I
was proved through the transformation of anti-14 (Scheme
7). The anti-14 was obtained through kinetic resolution
of 12. The transformation of anti-14 to 9 was carried out
under the same reaction conditions as described in
1
Scheme 4 for 14. Both H NMR and 19F NMR spectra of
Th e Ster eosp ecific Con ver sion of 17 to th e F u r a -
n ose 9 w ith Ar a bin o Con figu r a tion a t th e C2 P osi-
tion . As described in Scheme 4, the C2 (-OBn) config-
uration of 17 was stereospecifically transformed into a
single arabino configuration in compound 9. To further
verify this stereospecific transformation, a series of
experiments were examined. It would be expected that
an optically pure product could be obtained in the ring
opening of furanose 9 if it was a 1/1 mixture of R/â
anomers. In addition, in our previous paper,25 acetylation
of lactol with acetic anhydride would give almost a single
anomer at room temperature, when the C2 position at
the lactol was substituted with a large group in a single
configuration. This is due to the assitance of the neigh-
boring group participation. Accordingly, treatment of
furanose 9 with sodium borohydride in methanol gave
the diol 18 in quantitative yield (Scheme 5). Both 19F
9 synthesized from anti-14 were identical to those of
compound 9 prepared from 14 (Scheme 4). It should be
mentioned that 1H NMR and 19F NMR spectra were
determined before column chromatography so that any
products could not be lost. These results suggested that
the S-configuration of the C2 position at intermediate II
should turn into the R-configuration and the furanose 9
was formed from the intermediate I.
Syn th esis of N1-(3-Deoxy-3,3-d iflu or o-D-a r a bin o-
fu r a n osyl)cytosin e 6. The Vorbruggen glycosylation
reaction of pyrimidines with peracylated ribose usually
proceeds efficiently and stereoselectively at room tem-
perature to yield the â-anomer.26 However, the glycosy-
lation of silylated N4-benzoylcytosine with 19 in the
presence of either trifluoromethylsilyl trifluoromethane-
sulfonate or SnCl4 at room temperature failed to give any
protected N1-(3-deoxy-3,3-difluoro-D-arabinofuranosyl)-
cytosine 20. Fortunately, when the reaction temperature
rose to 80 °C, the nucleoside 20 was isolated in 70% yield
as a mixture of two isomers (20a /20b ) 1/1, Scheme 8)
along with the recovery of 19 (9%). Compounds 20a and
20b could be resolved by column chromatography. This
may be explained in terms of the so-called armed-
diarmed principle reported by Mootoo et al.27 Deprotec-
1
NMR and H NMR spectra of the crude reaction mixture
showed that only one compound was produced. At the
same time, as it would be expected, 3-deoxy-3,3-difluoro-
D-arabinofuranose 9 was subjected to acetic anhydride
at room temperature, the R anomer 19 was obtained
nearly as a single product (Scheme 5, R/â ) 17.0/1
determined by 19F NMR before column, 94% yield). These
results further confirmed that the C2 (-OBn) in com-
pound 9 had the arabino configuration.
(26) (a) Niedballa, U.; Vorbruggen, H. J . Org. Chem. 1974, 39, 3654.
(b) Niedballa, U.; Vorbruggen, H. J . Org. Chem. 1976, 41, 2084. (c)
Vorbrugen, H. Krolikiewicz, K,; Bennua, B. Chem. Ber. 1981, 114, 1234.
(d) Vorbruggen, H.; Hofle, G. Chem. Ber. 1981, 114, 1256.
(27) Mootoo, D. R.; Konradsson, P.; Udodong, U.; Fraser-Reid, B. J .
Am. Chem. Soc. 1988, 110, 5583.
The proposed mechanism of this stereospecific trans-
formation was described in Scheme 6. The aldehyde
(25) Zhang, X.; Qing, F.-L.; Yu, Y. J . Org. Chem. 2000, 65, 7075.
9030 J . Org. Chem., Vol. 68, No. 23, 2003

3-Deoxy-3,3-difluoro-D-arabinofuranose
SCHEME 6. P r op osed Mech a n ism of th e Ster eosp ecific Con ver sion of Com p ou n d 17 to F u r a n ose 9 w ith
Ar a bin o Con figu r a tion a t th e C2 P osition
SCHEME 7. Syn th esis of 9 fr om a n ti-14
SCHEME 8. Syn th esis of N1-(3-Deoxy-3,3-d iflu or o-
tion of the benzoyl of 20a and 20b in the presence of a
D-a r a bin ofu r a n osyl)cytosin e 6a
saturated solution of ammonia in methanol gave the C2′
position benzyl-protected R-cytidine 21a in 92% yield and
â-cytidine 21b in 88% yield, respectively. The 2′-benzyl
group of 21a was removed with hydrogen and 20% Pd-
(OH)2 on charcoal in methanol and cyclohexane at room
temperature for 12 h.28 It was found that over-hydroge-
nated compound 22 was formed in 76% yield. When the
reaction time was shortened to 2.5 h (the starting
material just disappeared), 49% of the N1-(3-deoxy-3,
3-difluoro-R-D-arabinofuranosyl)cytosine 23 was obtained
along with 10% of over-hydrogenated product 22. At-
tempts to improve the yield of 23 under other hydrogena-
tion conditions failed. For example, treatment of 21a with
hydrogen in the presence of Pd/C in THF at room
temperature resulted in the cleavage of the glycosyl bond.
To our delight, hydrogenation of 21b under the catalysis
of 20% Pd(OH)2 on charcoal in methanol and cyclohexane
afforded our target molecule N1-(3-deoxy-3,3-difluoro-â-
D-arabinofuranosyl)cytosine 6 in 83% yield. This indi-
cated that protected cytidine 21a was more reactive than
21b.
The configuration of the anomeric center was assigned
mainly by 1H NMR, in which the anomer with H4′ at
a
Key: (a) N4-benzoylcytosine, N,O-bis(trimethylsilyl)acetamide,
lower field was assigned as the R anomer and the one at
higher field was assigned as the â anomer on the basis
of the deshielding effect of the base moiety.29 This
TMSOTf, CH3CN, from 0 to 80 °C, 36 h; (b) saturated NH3/MeOH,
rt, 12 h; (c) H2 (1 atm), 20% of Pd(OH)2 on charcoal, MeOH/
cyclohexane (3:1), rt, 2.5 h.
assignment was further confirmed by the X-ray crystal-
(28) Tippie, M. A.; Martin, J . C.; Smee, D. F.; Matthews, T. R.;
Verheyden, J . P. H. Nucleosides Nucleotides 1984, 3, 525.
lography of 6 (Figure 2).
J . Org. Chem, Vol. 68, No. 23, 2003 9031

Zhang et al.
SCHEME 9. Syn th esis of Diflu or oth ion u cleosid es
29 via th e P u m m er er Rea ction a
SCHEME 10. Syn th esis of Diflu or oth ion u cleosid es
34 via th e P u m m er er Rea ction a
a
Key: (a) (i) PhOC(S)Cl, DMAP, CH3CN, from 0 °C to rt, 12 h,
(ii) Bu3SnH, AIBN, toluene, 80 °C, 40 min; (b) m-CPBA, CH2Cl2,
-40 °C, 40 min; (c) silylated N4-benzoylcytosine, TMSOTf, DCE,
0 °C, 30 min; (d) saturated NH3/MeOH, rt, 12 h.
a
Key: (a) (i) MsCl, Py/CH2Cl2, from 0 °C to rt, 12 h; (ii)
Na2S‚9H2O, DMF, 100 °C, 1 h; (b) BCl3 (1 M in CH2Cl2), MeOH,
satd aq NaHCO3, -70 °C, 3 h; (c) Bz2O, Et3N, DMAP, CH3CN, rt,
2 h; (d) m-CPBA, CH2Cl2, -40 °C, 40 min; (e) silylated N4-
benzoylcytosine, TMSOTf, DCE, 0 °C, 50 min; (f) satd NH3/MeOH,
rt., 12 h.
Attem p t To P r ep a r e L-3′-Deoxy-3′,3′-d iflu or o- 4′-
th iocytid in e 7. As shown in Scheme 1, 4′-thiofuranose
8 was a key precursor for the synthesis of 7. To synthesize
the 4′ -thiofuranose derivatives, the methods reported by
Dyson et al. 30 and Leydier et al.31 seemed to be the most
promising. The synthetic tactics involved two consecutive
SN2 reactions of D-ribose. Accordingly, compound 8 was
prepared as shown in Scheme 9. Mesylation at the C1
and C4 positions of 18, followed by treatment with
sodium sulfide in DMF, resulted in a ring closure to
thiofuranose 8 with a single configuration isomer in 69%
yield for the two steps with inversion of the stereochem-
istry at C4 position. The absolute configuration of com-
pound 8 was determined by X-ray crystal structure
(Figure 3). Treatment of 8 with boron trichloride in
methylene chloride followed by quenching with methanol
and saturated aqueous sodium bicarbonate afforded
compound 24 in quantitative yield. The alcohol 24 was
reacted with benzoic anhydride, triethylamine, and cata-
lytic DMAP in CH3CN to give benzoate 25 in quantitative
yield. Surprisingly, the oxidation of 25 with m-CPBA
followed by condensation with silylated N4-benzoylcy-
tosine failed to give our desired protected nucleosides 27′,
but the regioisomer 27 was obtained in 47% yield (two
steps) along with â-elimination product 28 in 22% yield
(two steps). The 19F NMR spectrum of compound 27
F IGUR E 4. ORTEP drawing of the X-ray crystallographic
structure of 34.
showed it was an epimeric mixture and the epimeric ratio
was 1:1. These two epimeric isomers could not be
separated by column chromatography. The removal of
benzoyl groups of 27 with a saturated solution of am-
monia in methanol provided the separated nucleosides
29a (41% yield) and 29b (42% yield). The steric assign-
ments of the final compounds were confirmed by the
NOESY experiments of 29a and 29b (Figure 3).
To further confirm the structure of nucleosides 29a ,b
and the regiochemistry of Pummerer reaction of the gem-
difluorinated thiofuranose, the Pummerer reaction of
gem-difluorinated thiofuranose 30 was investigated. Con-
version of 24 to the phenoxythiocarbonyl derivative
followed by the radical deoxygenation of the latter gave
compound 30 in 80% yield (Scheme 10). Interestingly,
oxidation of 30 with m-CPBA followed by the condensa-
tion with silylated N4-benzoylcytosine also gave the
racemic nucleoside 32 (38% yield, two steps) and â-elim-
ination product 33. Deprotection of compound 32 was
accomplished by ammonolysis to give nucleoside 34 in
85% yield. The structure of 34 was determined by 1H
NMR spectroscopy and further confirmed by X-ray crys-
tallography (Figure 4).
(29) Chun, B. K.; Olegen, S.; Hong, J . H.; Newton, M. G.; Chu, C.
K. J . Org. Chem. 2000, 65, 685.
(30) Dyson, M. R.; Coe, P. L.; Walker, R. T. Carbohydr. Res. 1991,
216, 237.
(31) Leydier, C.; Bellon, L.; Barascut, J .-L.; Deydier, J .; Maury, G.;
Pelicano, H.; Alaoui, M. A. E.; Imbach, J .-L. Nucleosides Nucleotides
1994, 13, 2035.
To date, numerous investigations have been carried out
to clarify the mechanism of the Pummerer reaction.32
Depending on the structures of the starting sulfoxides,
9032 J . Org. Chem., Vol. 68, No. 23, 2003

3-Deoxy-3,3-difluoro-D-arabinofuranose
SCHEME 11. P r op osed Mech a n ism of th e
P u m m er er Rea ction of gem -Diflu or in a ted
Su lfoxid es 26 a n d 31
proton is preferable to be removed. Thus, the presence
of difluoromethylene in our reaction facilitated proton
removal of 35 to produce trimethylsilylsulfonium ylide
36 via path a instead of path b, while at the same time
retarding departure of the leaving group due to strong
inductive effects of difluoromethylene. Although the steric
effect is a major factor in the regiochemistry of the
Pummerer reaction, it is less important than the acidity
of the R-proton in our reaction. This ylide subsequently
underwent concurrent electron reorganization and S-O
bond cleavage to give the ion pair 38. This step was rate
determining in the reaction in which proton removal was
fast and reversible and S-O bond cleavage was slow. In
the final step, condensation of silylated N4-benzoylcy-
tosine with 39 gave the regioisomer nucleoside 27 or 32.
But in this step, two pathways, nucleophilic attack (path
c) and â- elimination (path d), would occur. These two
paths competed with each other, and path c might be
more preferable than path d.
In summary, we have developed a stereoselctive and
efficient method to synthesize 3-deoxy-3,3-difluoro-D-
arabinofuranose 9, which was proved to be a useful
intermediate for the synthesis of difluoronucleoside
analogues. At the same time, we studied the epimeriza-
tion of compound 14 and found that the benzylation of
alcohol 12 with less than 1.0 equiv of sodium hydride
could kinetically resolve compound anti-14. In addition,
the stereospecific conversion of the C2 (-OBn) configu-
ration of 17 to a single arabino configuration in compound
9 was studied. The furanose 9 with the arabino config-
uration at the C2 position was thought to be formed from
aldehyde I, since the intermediate II could be easily
transformed into I via enolate III.
With the difluorinated furanose 9 as the key interme-
diate, the first synthesis of N1-(3-deoxy-3,3-difluoro-â-D-
arabinofuranosyl)cytosine 6 was accomplished in 10
steps. 4′-Thiofuranose 25 was easily synthesized from 9.
The oxidation of 25 followed by the condensation with
silylated N4-benzoylcytosine (Pummerer reaction) failed
to give our desired protected nucleoside L-3′-deoxy-3′,3′-
difluoro-4′-thiocytidine 27′, but the regioisomer 27 was
obtained. The regiochemistry of the Pummerer reaction
was determined by the kinetic acidity of the R-proton of
4′-thiofuranose 25.
both E1cb and E2 pathways are known to give the
R-thiocarbocation intermediates. In our reaction, the
E1cb mechanism could be used to explain the formation
of the unexpected nucleosides. As illustrated in Scheme
11, trimethylsilylation of the sulfoxide by TMSOTf to give
35 was the initial step of the Pummerer reaction.31d In
the following step, two properly positioned protons could
be removed, so there were two pathways in this step. It
is well-known that Elcb Pummerer reactions are observed
when the substituted group at the â-position of sulfonium
salt is an electron-withdrawing group and the more acidic
Ack n ow led gm en t. We thank the National Natural
Science Foundation of China and Shanghai Municipal
Scientific Committee for funding this work.
Su p p or tin g In for m a tion Ava ila ble: Experimental pro-
cedures and analytical data for all new compounds and
crystallographic data for compounds 6, 8, and 34 (CIF). This
material is available free of charge via the Internet at
http://pubs.acs.org.
(32) (a) J ohnson, C. R.; Sharp, J . C.; Phillips, W. G. Tetrahedron
Lett. 1967, 5299. (b) Davenport, D. A.; Moss, D. B.; Rhodes J . E.; Walsh,
J . A. J . Org. Chem. 1969, 34, 3353. (c) J ones, D. N.; Helmy, E.;
Whitehouse, R. D. J . Chem. Soc., Perkin Trans. 1 1972, 1329. (d)
Grierson, D. S.; Husson, H.-P. In Comprehensive Organic Synthesis;
Trost, B. M., Fleming, I., Ed.; Pergamon Press: Oxford, 1991; Vol. 6,
pp 909-947 and references therein.
J O034512I
J . Org. Chem, Vol. 68, No. 23, 2003 9033