Synthesis of L-β-3′-deoxy-3′,3′-difluoro-4′- thionucleosides

Article abstract of DOI:10.1021/ol062576k

(Chemical Equation Presented) An efficient route to L-β-3′- deoxy-3′,3′-difluoro-4′-thionucleosides, thio-containing analogues of highly bioactive gemcitabine, is described. Our synthesis highlighted the installation of the thioacetyl group in high effici

Full text of DOI:10.1021/ol062576k

ORGANIC
LETTERS
2006
Vol. 8, No. 26
6083-6086
Synthesis of
-3 -Deoxy-3′,3′-difluoro-4′-thionucleosides
L
-â
′
Feng Zheng, Xiao-Hua Zhang, Xiao-Long Qiu, Xingang Zhang, and
Feng-Ling Qing*
Key Laboratory of Organofluorine Chemistry, Shanghai Institute of Organic
Chemistry, Chinese Academy of Sciences, 354 Fenglin Lu, Shanghai 200032, China
flq@mail.sioc.ac.cn
Received October 19, 2006
ABSTRACT
An efficient route to L-â-3′-deoxy-3′,3′-difluoro-4′-thionucleosides, thio-containing analogues of highly bioactive gemcitabine, is described.
Our synthesis highlighted the installation of the thioacetyl group in high efficiency and construction of 3-deoxy-3,3-difluorothiofuranose skeleton
in a novel method.
gem-Difluoromethylene (CF2) has been suggested by Black-
burn as an isopolar and isosteric substituent for oxygen.1
Since then, the CF2 group was extensively used to modify
nucleoside analogues. For example, 2′-deoxy-2′,2′-difluoro-
cytidine (gemcitabine 1, Figure 1) has been approved as a
drug for solid tumor treatment.2 Unnatural L-configuration
nucleosides have drawn considerable attention due to their
lower host toxicity while maintaining good activity in
comparison with their corresponding D-nucleosides.3 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.4
Recently, a number of D-(L-)difluoromethylene-containing
thionucleosides have been prepared, such as D-2′-deoxy-2′,2′-
difluoro-4′-thiocytidine 2,5 L-2′-deoxy-2′,2′-difluoro-4′-thio-
cytidine 3,6 and D-2′,3′-dideoxy-3′,3′-difluoro-4′-thiocytidine
4.7 Based on the above considerations and our efforts to
prepare the fluorinated sugar nucleosides,8 we designed our
target molecules L-3′-deoxy-3′,3′-difluoro-4′-thionucleosides
(1) Blackburn, C. M.; England, D. A.; Kolkmann, F. J. Chem. Soc., Chem.
Commun. 1981, 930.
(2) (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.
(3) For a minireview, see: Mathe, C.; Gosselin, G. AntiViral Res. 2006,
Figure 1. Rationale for the design of the target molecule 5.
71, 276.
(4) For review; see: Yokoyama, M. Synthesis 2000, 1637.
10.1021/ol062576k CCC: $33.50
© 2006 American Chemical Society
Published on Web 11/30/2006

5 (Figure 1). The gem-difluoromethylene (CF2) group was
at the C3′ position of compounds 5. Moreover, nucleosides
5 would combine the characteristics of compounds 1-4
based on the bioisosteric rationale (Figure 1).
Scheme 2
Our previous attempt to prepare L-3′-deoxy-3′,3′-difluoro-
4′-thiocytosine 5 via Pummerer reaction of compound 6
failed (Scheme 1).9 The oxidation of 6 followed by the
Scheme 1
ring opening of 13 with BnSH afforded the 1-thiofuranoside
14 as the main product, and our desired thioacetal 15 was
provided only in very low yield. More unfortunately, the
corresponding triflate 16, obtained via O-triflation of 15,
decomposed very quickly at room temperature.
Then, another strategy to thiosugar, pioneered by Mont-
gomery et al.,10 was attempted. Their method highlighted
the fact that reductive deprotection of SAc to SH with
DIBAL-H and simultaneous reduction of the methyl ester
to -CHO in the same molecule would give rise to thioactol
formation via spontaneous cyclization. Thus, developing a
practical route to thioacetate 19 starting from compound 11
was what should be first addressed (Table 1).
condensation with silyated N4-benzoylcytosine (Pummerer
recation) did not give our desired protected nucleoside 7,
but the regioisomer 8. The regiochemistry of the Pummerer
recation was determined by the kinetic acidity of the R-proton
of 4′-thiofuranose 6.
Table 1. Preparation of the Thioacetate 19
The new synthetic route was based on the supposition that
the target molecules 5 could be derived from the precursor
9 by introduction a base moiety using the glycosylation
reactions (Scheme 1). Intermediate 9 would be reached
through ring closure of the our developed versatile gem-
difluorinated synthon 10, which was easily prepared accord-
ing to our reported methodology.8a,9
entry
R
conditions
19 (%)
NRa
1
2
Ms AcSH (3 equiv)/CsF (3 equiv)/DMF/rt
Ms AcSH (3 equiv)/CsF (3 equiv)/DMF/50 °C defluorin
ationa
The construction of the special skeleton 9 appeared to be
a great challenge. Recently, Chu group reported the synthesis
of thiosugar through the corresponding riboside,7 which
stimulated us to prepare the intermediate 9 using the same
strategy. Thus, 3-deoxy-3,3-difluoro-D-arabinofuranose 12,
prepared from optically pure gem-difluorinated synthon 11
via our reported procedure,9 was converted to compound 13
by treatment with TsOH/MeOH (Scheme 2). However, the
3
Tf KSAc (3 equiv)/DMF
defluorin
ationa
72b
4
5
Tf AcSH (3 equiv)/CsF (3 equiv)/DMF/rt
Tf AcSH (5.4 equiv)/CsF (5.4 equiv)/DMF/rt 86b
a Determined by 19F NMR spectra. b Yield was based on chromatography
isolation over silica gel.
Otera and co-workers reported an efficient method of CsF/
DMF-mediated nucleophilic inversion of secondary mesyl-
ates and tosylates.11 Their strategy could be successfully
ultilized with a variety of oxygen-, sulfur-, nitrogen- and
carbo-nucleophiles. Thus, we think that thioacetate 19 could
(5) Yohimura, Y.; Kitano, K.; Yamada, K.; Satoh, H.; Watanabe, M.;
Miura, S.; Sakata, S.; Sasaki, T.; Matsuda, A. J. Org. Chem. 1997, 62, 3140.
(6) Jeong, L. S.; Moon, H. R.; Choi, Y. J.; Chun, M. W.; Kim, H. O. J.
Org. Chem. 1998, 63, 4821.
(7) Zhu, W.; Chong, Y.; Choo, H.; Mathews, J.; Schinazi, R. F.; Chu,
C. K. J. Med. Chem. 2004, 47, 1631.
(8) (a) Xu, X. H.; Qiu, X.-L.; Zhang, X.; Qing, F.-L. J. Org. Chem.
2006, 71, 2820. (b) Meng, W. D.; Qing, F.-L. Curr. Top. Med. Chem. 2006,
6, 1499.
(9) Zhang, X.; Xia, H.; Dong, X.; Jin, J.; Meng, W. D.; Qing, F.-L. J.
Org. Chem. 2003, 68, 9026.
(10) Secrist, J. A., III; Riggs, R. M.; Tiwari, K. N.; Montgomery, J. A.
J. Med. Chem. 1992, 35, 533.
(11) Otera, J.; Nakazawa, K.; Sekoguchi, K.; Orita, A. Tetrahedron 1997,
53, 13633.
6084
Org. Lett., Vol. 8, No. 26, 2006

be provided by treatment of the corresponding mesylate 17,
obtained via exposure of compound 11 to MsCl/pyridine,
with AcSH under a CsF/DMF system. However, no reaction
occurred when compound 17 was treated with AcSH (3
equiv)/CsF (3 equiv) in DMF at room temperature (Table 1,
entry 1). When the reaction was performed at 50 °C, 19F
NMR of reaction mixture showed that the defluorination
occurred (Table 1, entry 2). In our opinion, failure of above
reactions was attributed to the facts that (1) a strong electron-
withdrawing gem-difluoromethylenyl group (CF2) could
effectively stabilize the neighboring C-O bond, which made
the mesyl group hard to remove, and (2) a strong electron-
withdrawing CF2 would enhance the acidity of 2-H in
compound 17, and thus, when reaction temperature was
increased, the dehydrofluorination reaction and other side
reactions occurred in the presence of Lewis base F-. Taking
the above analyses together, we envisioned that the triflo-
romethanesulfonyl as the leaving group should probably
afford our desired compound. Thus, triflate 18 was first
prepared in good yield via O-triflation of 11. To our delight,
although treatment of compound 18 under the normal condi-
tions (KSAc/DMF) still resulted in dehydrofluorination (Table
1, entry 3), substitution of AcSH (3 equiv)/CsF (3 equiv)
for KSAc smoothly furnished our desired thioacetate 19 in
72% yield (Table 1, entry 4), which could be further
improved to 86% when 5.4 equiv of AcSH/CsF was used
(Table 1, entry 5).
Scheme 4
preparing the thiofuranose 20 from 19. What surprised us
was that only compound 20 was isolated, which showed that
the stereospecific transformation of the C3 (-OBn) config-
uration in compound 21 also happened. It was noteworthy
that the C3 (-OBn) configuration of compound 20 was
contrary to that of compound 12.
The proposed mechanism of this stereospecific transfor-
mation is outlined in Scheme 5. The aldehyde intermediates
Scheme 5
With the thioacetate 19 in hand, we then focused on the
construction of the thiofuranose 9. Fortunately, we were
delighted to find that the thiofuranose 20 could be provided
in 64% overall yield (Scheme 3) when the resultant aldehyde,
Scheme 3
in situ obtained via hydrolysis of the thioacetate 19 with TFA
and oxidation of the resulting diols with NaIO4, was treated
with acidic methanol (1 M HCl) at room temperature for 12
h. The isomers of compound 20 could not be separated on
silica gel chromatography.
During the synthesis of N1-(3-deoxy-3,3-difluoro-D-ara-
binofuranosyl)cytosine,9 we accidently found that the C3
(-OBn) configuration in compound 10 was stereospecifically
transformed into a single arabino configuration in 12
(Scheme 4). As shown in Scheme 4, the hydrogen-bond-
mediated four-membered ring may play an important role
in the favorite transition state of the cyclization to form
compound 12. We were interested in investigating whether
this stereospecifical transformation also occurred in the
cyclization of the thioacetate. Thus, compound 21 (a mixture
of C3 epimers) was first prepared starting from compound
10 using the aforementioned starategy. Then, compound 21
was subjected to the same conditions as described above for
I and II could be formed from compound 21 and were in
tautomeric equilibrium through the intemediate III. The
stereospecific formation of compound 20 from intermediate
I with the S configuration at the C2 position was thought to
be faster than the formation of compound 20′ because a huge
gauche between the BnO group and the BzOCH2 group in
intermediate II existed. In addition, the attack of the free
hydroxyl group of I on the carbonyl group from both the re
and si faces of the aldehyde moiety resulted in a 1/1 mixture
of R/â anomers of 20.
The acetylation of 20 with acetic anhydride in CH2Cl2
afforded the â anomer 22 as the main product (R/â ) 1/6)
Org. Lett., Vol. 8, No. 26, 2006
6085

Finally, removal of the benzoyl groups in compounds 25-
27 with saturated methanolic ammonia smoothly produced
the desired free nucleosides 28-30.
Scheme 6
The stereochemistry of compound 26 has been established
by 2D NMR NOESY experiments. As shown in Figure 2,
Figure 2. NOE correlation of compound 26 and X-ray structure
of compound 28.
(Scheme 6), which resulted from the assistance of neighbor-
ing large group participation (Scheme 6).12 In view of the
fact that the glycosylation of furanose possessing an acyloxy
group in C-2′ position would involve the oxonium intermedi-
ate, which might induce the attack of the silylated base from
the contrary face of C-2′ acyloxy group13 to give the â
anomer nucleosides as main product and for the conveniency
of the following removal of the protecting groups, we decided
to replace the benzyloxy group in compound 22 with a
benzoyloxy group. Thus, treatment of the acetate 22 with
NaBrO3/Na2S2O414 smoothly provided the alcohol 23 in 87%
yield, which was then benzoylated with Bz2O/DMAP/Et3N
to produce the key intermediate 24 in 92% yield. Then,
coupling of compound 24 with various persilylated pyrim-
idines were performed in refluxed acetonitrile under Vor-
bru¨ggen conditions.15 As expected, all the reactions gave the
â anomers as the main products. Trace of the R anomers
was detected by 19F NMR spectra of the crude reaction
mixture, but was not isolated by flash chromatography.
correlation between H1′ and H4′ was clearly observed in this
compounds. In addition, the H2′ proton strongly correlated
with H6 protons of base. All of these showed that the
nucleoside derivative 26 is â anomer. Furthermore, the
structure of compound 28 was confirmed by X-ray crystal
analysis (Figure 2).
In conclusion, we have designed and synthesized the new
L-3′-dideoxy-3′,3′-difluoro-4′-thionucleosides 28-30 based
on our versatile gem-difluorinated synthon. Our synthesis
highlighted the installation of thioacetyl group in high
efficiency and construction of thiofuranose skeleton in a
novel method. Antiviral and cytotoxicity evaluations of
herein synthesized L-3′-deoxy-3′,3′-difluoro-4′-thionucleo-
sides are currently in progress and will be reported elsewhere.
Acknowledgment. The National Natural Science Foun-
dation of China and Shanghai Municipal Scientific Com-
mittee are greatly acknowledged for funding this work.
(12) Zhang, X.-G.; Qing, F.-L.; Yu, Y.-H. J. Org. Chem. 2000, 65, 7075.
(13) Bio, M. M.; Xu, F.; Waters, M.; Williams, J. M.; Savary, K. A.;
Cowden, C. J.; Yang, C.-H.; Buck, E.; Song, Z. J.; Tschaen, D. M.; Volante,
R. P.; Reamer, R. A.; Grabowski, E. J. J. J. Org. Chem. 2004, 69, 6257.
(14) Adinolfi, M.; Barone, G.; Guariniello, L.; Iadonisi, A. Tetrahedron
Lett. 1999, 40, 8439.
Supporting Information Available: Detailed experi-
mental procedures and analytical data for all new compounds
and crystallographic data for compound 28. This material is
available free of charge via the Internet at http://pubs.acs.org.
(15) Vorbru¨ggen, H.; Krolikiewicz, K.; Bennua, B. Chem. Ber. 1981,
114, 1234.
OL062576K
6086
Org. Lett., Vol. 8, No. 26, 2006