An Alternative Synthesis of the Antineoplastic Nucleoside 4′-ThioFAC and Its Application to the Synthesis of 4′-ThioFAG and 4′-Thiocytarazid

Article abstract of DOI:10.1021/jo990958g

Previously, we synthesized 4′-thioFAC, a novel antineoplastic cytosine nucleoside, by developing an original method. However, several problems remained. To overcome these problems, we have developed an alternative method for the synthesis of 4′-thionucleosides. In the original synthesis, carbons from C1 to C5 of D-glucose were used. The new method also starts from D-glucose but uses carbons closer to the tail (C2-C6). A dibenzoyl derivative obtained by this approach was brominated at the anomeric position to give a 1-bromide derivative. Fusion of the 1-bromide and persilylated acetylcytosine, followed by deprotection, predominantly gave a β-anomer of 4′-thioFAC. The reaction of 2,6-diaminopurine with the 1-bromide in the presence of TMS triflate gave a glycosylated product in good yield. After deprotection, the resulting 1:1 anomeric mixture of free nucleosides was treated with adenosine deaminase to give a β-anomer of 4′-thioFAG, a guanine congener of 4′-thioFAC, selectively. Using a similar approach, we synthesized 4′-thiocytarazid, which was not possible using the original method.

Full text of DOI:10.1021/jo990958g

7912
J . Org. Chem. 1999, 64, 7912-7920
An Alter n a tive Syn th esis of th e An tin eop la stic Nu cleosid e
4′-Th ioF AC a n d Its Ap p lica tion to th e Syn th esis of 4′-Th ioF AG a n d
4′-Th iocyta r a zid
Yuichi Yoshimura,*,† Mikari Endo, Shinji Miura, and Shinji Sakata
Biochemicals Division, Yamasa Corporation, 2-10-1 Araoicho, Choshi, Chiba 288-0056, J apan
Received J une 14, 1999
Previously, we synthesized 4′-thioFAC, a novel antineoplastic cytosine nucleoside, by developing
an original method. However, several problems remained. To overcome these problems, we have
developed an alternative method for the synthesis of 4′-thionucleosides. In the original synthesis,
carbons from C1 to C5 of D-glucose were used. The new method also starts from D-glucose but uses
carbons closer to the tail (C2-C6). A dibenzoyl derivative obtained by this approach was brominated
at the anomeric position to give a 1-bromide derivative. Fusion of the 1-bromide and persilylated
acetylcytosine, followed by deprotection, predominantly gave a â-anomer of 4′-thioFAC. The reaction
of 2,6-diaminopurine with the 1-bromide in the presence of TMS triflate gave a glycosylated product
in good yield. After deprotection, the resulting 1:1 anomeric mixture of free nucleosides was treated
with adenosine deaminase to give a â-anomer of 4′-thioFAG, a guanine congener of 4′-thioFAC,
selectively. Using a similar approach, we synthesized 4′-thiocytarazid, which was not possible using
the original method.
Ch a r t 1
In tr od u ction
The discovery of the potent antiherpes virus activity
of 2′-deoxy-4′-thionucleosides has stimulated the develop-
ment of 4′-thionucleosides as potential antiviral agents.1,2
The resistance of the glycosidic bond of 4′-thionucleosides
to enzymatic hydrolysis catalyzed by nucleoside phos-
phorylase3 is one of the critical points in nucleoside
antiviral therapeutics. Several 4′-thionucleosides have
been reported to be active against herpes viruses.1,2,4,5 In
addition, some have also shown antiretrovirus activity:
e.g., L-2′,3′-dideoxy-didehydro-4′-thiocytidine (L-D4C, 1,
Chart 1) has potent antihuman immunodeficiency virus
activity.6 Another important feature of 4′-thionucleosides
is their antineoplastic activity.1,7 Recently, our attention
has been focused on the synthesis and antineoplastic
activity of 2′-substituted-4′-thiocytidine analogues.8,9
Among them, 1-(2-deoxy-2-fluoro-â-D-4-thio-arabino-
pentofuranosyl)cytosine (4′-thioFAC, 2) is noteworthy.9
* Corresponding author: Yuichi Yoshimura. Tel: +81-3-3784-8187.
FAX: +81-3-3784-8252. E-mail: yuyoshim@pharm.showa-u.ac.jp.
† Present address: School of Pharmaceutical Sciences, Showa
University, 1-5-8 Hatanodai, Shinagawa-ku, Tokyo 142-8555, J apan.
(1) Secrist, J . A.; Tiwari, K. N.; Riordan, J . M.; Montgomery, J . A.
J . Med. Chem. 1991, 34, 2361-2366.
(2) (a) Dyson, M. R.; Coe, P. L.; Walker, R. T. J . Chem. Soc., Chem.
Commun. 1991, 741-742. (b) Dyson, M. R.; Coe, P. L.; Walker, R. T.
J . Med. Chem. 1991, 34, 2782-2786.
4′-ThioFAC has prominent and broad antitumor activities
against various human solid tumor cell lines in vitro as
well as in vivo.9,10 Furthermore, 4′-thioFAC is an orally
active antitumor agent; it has shown therapeutic effects
on nude mice bearing colon and stomach cancers when
given orally.10
Despite the unique biological activities of 4′-thionucleo-
sides, the difficulty of synthesizing 4′-thionucleosides has
impeded the investigation of structure-activity relation-
ships (SAR). Thus, the production of new 4′-thionucleo-
side analogues has often depended upon the development
of new synthetic methods.8 In fact, for the synthesis of
2′-substituted-4′-thiocytidine, we exploited a novel syn-
thetic method, which was successfully used to synthesize
(3) 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 Chemotherapeutic Agents; Sartorelli, A. C., Lazo, J . S., Bertino,
J . R., Eds.; Academic Press: New York, 1981; pp 229-252.
(4) Yoshimura, Y.; Watanabe, M.; Satoh, H.; Ashida, N.; Ijichi, K.;
Sakata, S.; Machida, H.; Matsuda, A. J . Med. Chem. 1997, 40, 2177-
2183.
(5) Satoh, H.; Yoshimura, Y.; Watanabe, M.; Ashida, N.; Ijichi, K.;
Sakata, S.; Machida, H.; Matsuda, A. Nucleosides Nucleotides 1998,
17, 65-79.
(6) Young, R. J .; Shaw-Ponter, S.; Thomson, J . B.; Miller, J . A.;
Cumming, J . G.; Pugh, A. W.; Rider, P. Bioorg. Med. Chem. Lett. 1995,
5, 2599-2604.
(9) Yoshimura, Y.; Kitano, K.; Yamada, K.; Satoh, H.; Watanabe,
M.; Miura, S.; Sakata, S.; Sasaki, T.; Matsuda, A. J . Org. Chem. 1997,
62, 3140-3152.
(10) Miura, S.; Yoshimura, Y.; Endo, M.; Machida, H.; Matsuda, A.;
Tanaka, M.; Sasaki, T. Cancer Lett. 1998, 129, 103-110.
(7) Tiwari, K. N.; Secrist, J . A.; Montgomery, J . A. Nucleosides
Nucleotides 1994, 13, 1819-1828.
(8) Yoshimura, Y.; Kitano, K.; Satoh, H.; Watanabe, M.; Miura, S.;
Sakata, S.; Sasaki, T.; Matsuda, A. J . Org. Chem. 1996, 61, 822-823.
10.1021/jo990958g CCC: $18.00 © 1999 American Chemical Society
Published on Web 09/24/1999

Synthesis of 4′-Thionucleosides
J . Org. Chem., Vol. 64, No. 21, 1999 7913
4′-thioFAC.9 However, our original synthesis of 4′-thio-
FAC has not been optimized. Several problems make it
unsuitable for large-scale preparation. First, an expensive
silyl protecting group (tert-butyldiphenylsilyl group) and
difficult-to-handle reagents, such as DAST, BBr3, and
mcpba, were used.9 In addition, in the most serious
problem in the synthesis of 4′-thioFAC, the undesired
R-isomer was predominantly formed (R: â ) 2.5:1) and
had to be separated from the desired â-isomer by a
complicated purification method.9 To overcome these
drawbacks, the development of an alternative synthetic
method that can selectively produce â-4′-thioFAC is
strongly desired.
Sch em e 1
With regard to these considerations, the improved
synthesis of 1-(2-deoxy-2-fluoro-â-D-arabino-pentofur-
anosyl)cytosine (FAC, 3), a 4′-oxy counterpart of 4′-
thioFAC reported by Watanabe, was suggestive.11 In
addition, Wistler et al. achieved the synthesis of pyrimi-
dine 4′-thioarabino nucleosides by a similar approach.12
Thus, we sought to apply Watanabe’s method to the
synthesis of 4′-thioFAC. In this paper, we describe the
alternative synthesis of a 2-fluoro-4-thiosugar derivative,
and its stereoselective coupling with persilylated N4-
acetylcytosine, which led to 4′-thioFAC.13 By applying
this new method, we synthesized 1-(2-deoxy-2-fluoro-â-
D-4-thio-arabino-pentofuranosyl)guanine (4′-thioFAG, 4)
and 1-(2-azido-2-deoxy-â-D-4-thio-arabino-pentofurano-
syl)cytosine (4′-thiocytarazid, 5). The former, previously
synthesized by us, has potent antiherpes virus activities
in vitro and had to be prepared on a multigram scale for
the in vivo antiviral assays in our laboratory. The latter,
which has been previously designed as a potential
antitumor agent, could not be obtained by the original
method. We also describe the synthesis of both com-
pounds.
expanding these chemistries. To improve the poor â-se-
lectivity at the glycosylation step, we investigated the
nucleophilic substitution of a 1-R-bromide of a 4-thiosugar
derivative, in place of the Lewis-acid-catalyzed reaction,
in the alternative synthesis.
According to our plan, 1,2:5,6-di-O-isopropylidene-R-
D-allofuranose (6), which was easily accessible from 1,2:
5,6-di-O-isopropylidene-R-D-glucose (I) in two steps,15
should be fluorinated at the 3-position. Instead of the
original method reported by Watanabe,11 we used an
alternative method, which was developed by a Bristol-
Myers group,16 with slight modification. Compound 6 was
treated with sulfuryl chloride and imidazole to give
imidazoyl sulfate 7, which was fluorinated by treatment
with potassium fluoride in refluxing 2-methoxyethanol
to give 3-fluoro derivative 8 in 77% yield from 6. Notably,
the substitution reaction of 7 with a fluoro anion pro-
ceeded under conditions that were more gentle (125 °C
in 2-methoxyethanol) than the previous examples, in
which a higher reaction temperature (210 °C in acet-
amide11 or 160 °C in 2,3-butanediol16) was required. In
addition, the reaction of 7 did not require any additive,
such as HF, which was used in the original method of
the Bristol-Myers group.16 As expected, the 3-fluoro
derivative 8 consisted of a single stereoisomer, the
inverted stereochemistry at C3 of which was confirmed
by converting the final compound. The 3-fluoro derivative
8 was selectively deblocked at the 5,6-isopropylidene
group under acidic conditions. Selective benzoylation at
the primary hydroxyl group of the resulting diol gave
6-benzoate 9 in 82% yield from 8. Mesylation at the C5
position of 9, followed by treatment with sodium meth-
oxide, resulted in debenzoylation and the subsequent
formation of epoxide to give 5,6-epoxide 10 in 82% yield
with inversion of the stereochemistry at C5. Treatment
of the 5,6-epoxide 10 with thiourea in refluxing methanol
gave a 5,6-thiirane derivative 11, the C5 configuration
of which had to be controlled by introducing a thio
functional group, since it was to be a D-thiosugar. The
5,6-thiirane ring of 11 was cleaved by treatment with
Resu lts a n d Discu ssion
Alter n a tive Syn th esis of 4′-Th ioF AC. Scheme 1
shows the plan for the alternative synthesis of 2 in
comparison with the original synthesis. In the original
synthesis, the glucose carbons from C1 to C5 were used
for a 4-thiosugar intermediate II, the chiralities of C2,
C3, and C4 of which were transferred from C4, C3, and
C2 of glucose, respectively.9 The glycosidic bond of 2 was
formed via Pummerer rearrangement, followed by the
Lewis acid-catalyzed glycosylation reaction, which re-
sulted in the predominant formation of the undesired
R-anomer.9 Although the alternative method also begins
with D-glucose, it uses carbons closer to the tail (C2-
C6), as in the case of Watanabe’s FAC synthesis.11 Thus,
the three sets of chiral centers from C2 to C4 in the
intermediate III are transferred from C3 to C5 of
D-glucose, respectively. The syntheses of 5- and 4-thio-
sugar derivatives, which involve the introduction of a
sulfur unit at C5 of glucose and other sugars, have been
reported previously.14 However, there have been few
reports concerning the synthesis of 4′-thionucleosides by
(11) Reichman, U.; Watanabe, K. A.; Fox, J . J . Carbohydr. Res. 1975,
42, 233-240.
(12) (a) Whistler, R. L.; Nayak, U. G.; Perkins, A. W. J . Org. Chem.
1970, 35, 519-521. (b) Whistler, R. L.; Doner, L. W.; Nayak, U. G. J .
Org. Chem. 1971, 36, 108-110.
(13) A part of this work has appeared: Yoshimura, Y.; Endo, M.;
Sakata, S. Tetrahedron Lett. 1999, 40, 1937-1940.
(14) Yuasa, H.; Tamura, J .; Hashimoto, H. J . Chem. Soc., Perkin
Trans. 1 1990, 2763-2769.
(15) Baker, D. C.; Horton, D.; Tindall, C. G. Carbohydr. Res. 1972,
24, 192-197.
(16) Tann, C. H.; Brodfuehrer, P. R.; Brundidge, S. P.; Sapino, C.;
Howell, H. G. J . Org. Chem. 1985, 50, 3644-3647.

7914 J . Org. Chem., Vol. 64, No. 21, 1999
Yoshimura et al.
Sch em e 2
potassium acetate in refluxing acetic anhydride and
acetic acid (5:1) for 2 days,14 to give a diacetate derivative
12 in 73% yield. Conversion of the diacetate derivative
12 to a desired 2-fluoro-4-thiosugar 13 was achieved by
the following 4 steps: (1) selective acidic hydrolysis of
the isopropylidene group by treatment with 90% trifluo-
roacetic acid at 0 °C, (2) oxidative scission of the resulting
diol with sodium periodate, (3) deacetylation and subse-
quent cyclization to a thiofuranose by treatment with
acidic methanol, and (4) benzoylation of the free hydroxyl
groups by treatment with benzoyl chloride in pyridine.
By these procedures, the 2-fluoro-4-thiosugar 13 was
obtained in 58% yield from 12 (Scheme 2).
As mentioned above, the major drawback of the
original synthesis of 4′-thioFAC was the formation of the
undesired R-anomer over that of the biologically active
â-anomer. Therefore, in this alternative synthesis we had
to improve the â-stereoselectivity at the glycosylation
step. Considering our previous results4,8,9 and those of
others,1,2,17-19 the Lewis acid-catalyzed glycosylation reac-
tion of 4-thiosugars tends to give undesired R-anomers
as major products. Furthermore, the Lewis acid-catalyzed
reaction of a 2-fluoroarabinose derivative predominantly
forms an R-substituted product, due to the strong elec-
trostatic effect of a fluoro substituent, which impairs the
approach of nucleophiles from the â-side.20 Therefore, the
previous syntheses of FAC and related compounds were
achieved by the nucleophilic substitution of the 1-R-
bromo-2-fluoroarabinose derivative with nucleobases.11,20,21
Bromination of the anomeric position of the 2-fluoro-
arabinose derivative, under the usual conditions, selec-
tively gave the 1-R-bromide, which was applied to per-
silylated pyrimidines to give â-2-fluoroarabino pyrimidine
nucleosides exclusively.11,20,21 Thus, the condensation of
silylated cytosine and 1-bromo-4-thiosugar, which would
be derived from 13, is an attractive choice to overcome
the most serious problem.
Compound 13 was subjected to acetolysis of the 1-meth-
oxy group to give 1-acetate 14 (90% yield), which was
treated with HBr/acetic acid in dichloromethane to give
the corresponding 1-bromide 15. The instability of 15
made it difficult to confirm its structure and determine
the ratio of R- and â-anomers based on the 1H NMR
spectrum. As in the case of the 4-oxy derivative men-
tioned above, we assumed that the 1-bromide 15 might
largely be composed of an R-bromide, due to the stereo-
electronic effect of the 2-fluoro substituent. Thus, soon
after bromination of the 1-acetate 14, the resulting
1-bromide 15 was subjected to the glycosylation of
persilylated N4-acetylcytosine without any purification
or inspection.
In general, a 1-halogeno-sugar undergoes anomeriza-
tion in a polar solvent. When persilylated pyrimidines
are reacted with 1-R-halogeno sugar in such a solvent,
the anomerization causes an increase in the unfavorable
formation of R-nucleosides. Walker reported that â-2′-
deoxynucleoside could selectively form when the pro-
tected 1-chloro-2-deoxyribose reacted in chloroform.22
Howell et al. fully investigated the glycosylation of the
1-bromo-2-fluoroarabinose derivative by an SN2 displace-
ment,20 which was first reported by Watanabe and Fox
as mentioned above, to synthesize a series of antiviral
2′-fluoroarabino-nucleosides, and obtained results similar
to those reported by Walker.20 On the basis of these
results, our first attempt at the glycosylation reaction by
the condensation of 15 with persilylated N4-acetylcytosine
was performed in 1,2-dichloroethane, but poor results
were obtained. The reaction in refluxing 1,2-dichloro-
ethane gave only trace amounts of the glycosylated
product. Similarly, other conditions with less-polar sol-
vents such as dichloromethane and carbon tetrachloride
gave only trace amounts or none of the glycosylated
products (data not shown). However, the latter unreacted
mixture in carbon tetrachloride was subjected to another
(17) Leydier, C.; Bellon, L.; Barascut, J .-L.; Deydier, J .; Maury, G.;
Pelicano, H.; El Alaoui, M. A.; Imbach, J .-L. Nucleosides Nucleotides
1994, 13, 2035-2050.
(18) Uenishi, J .; Takahashi, K.; Motoyama, M.; Akashi, H. Chem.
Lett. 1993, 255-256.
(19) Hartsel, S. A.; Marshall, W. S. Tetrahedron Lett. 1998, 39, 205-
208.
(20) Howell, H. G.; Brodfuehrer, P. R.; Brundidge, S. P.; Benigni,
D. A.; Sapino, C. J . Org. Chem. 1988, 53, 85-88.
(21) (a) Watanabe, K. A.; Reichman, U.; Hirota, K.; Lopez, C.; Fox,
J . J . J . Med. Chem. 1979, 22, 21-24. (b) Watanabe, K. A.; Su, T.-L.;
Klein, R. S.; Chu, C. K.; Matsuda, A.; Chun, M. W.; Lopez, C.; Fox, J .
J . J . Med. Chem. 1983, 26, 152-156.
(22) Hubbard, A. J .; J ones, A. S.; Walker, R. T. Nucl. Acids Res.
1984, 12, 6827-6837.

Synthesis of 4′-Thionucleosides
J . Org. Chem., Vol. 64, No. 21, 1999 7915
Sch em e 3
Sch em e 4
4′-thioFAG. To construct a glycosidic linkage of 4′-
thioFAG, the SN2 displacement of 1-bromide 15 using
metal salts of purine bases seems to be promising, since
the stereoselective formation of â-nucleoside is expected.
Although we investigated the glycosylation reaction by
using metal salts of 2,6-dichloropurine or 2-amino-6-
chloropurine and 1-bromide 15 under various conditions,
none of the reactions was successful.
Therefore, improvement of the synthesis of 4′-thioFAG
mainly focused on increasing the yield of the glycosylation
reaction of 2,6-diaminopurine, which was catalyzed by a
Lewis acid. We used both 1-acetate 14 and 1-bromide 15
as glycosyl donors and investigated the glycosylation
reaction of 2,6-diaminopurine under various Lewis acid
conditions. Of the conditions we tested, the reaction of
1-bromide 15 in the presence of TMS triflate as a Lewis
acid in acetonitrile at 60 °C gave the best results, in
which the glycosylated product 19 was obtained in 75%
yield (R:â ) 1.2:1). The use of stannic chloride, instead
of TMS triflate, caused a decrease in the yield of 19.
Likewise, the reaction of 1-acetate 14 also resulted in a
decrease in 19 (data not shown). An anomeric mixture
of the resulting 19 was deprotected by treatment with
concentrated NH4OH/MeOH to give a mixture of R- and
â-anomers of 2,6-diamino derivative 20 in good yields.
To prepare 4, it was necessary to hydrolyze at the
6-amino group and separate the resulting biologically
active â-4 from the inactive R-anomer. This could be
achieved by using adenosine deaminase, which can
discriminate a natural â-form from an R-form around a
glycosidic bond. In fact, in a small-scale reaction of 20,
adenosine deaminase selectively recognized â-20.24 There-
fore, we could easily separate â-4′-thioFAG 4 from the
unhydrolyzed R-2,6-diaminopurine derivative R-20 by
simple reversed-phase column chromatography. How-
ever, the low water-solubility of 20 made it difficult to
reproduce the same reaction in a large-scale synthesis
in which the volume of reaction buffer had to be de-
creased to half of that in the small-scale reaction because
of the efficiency of column chromatography. Since this
slowed the reaction rate, the reaction time had to be
prolonged to complete the hydrolysis. As a result, un-
desired R-4′-thioFAG, which no longer could be separated
from 4, was also obtained. We overcame this problem by
stopping the hydrolysis at an appropriate time. After
unreacted â-20 and undesired R-20 were separated from
â-4, hydrolysis of the 6-amino group by adenosine deami-
nase was repeated. Overall, 4′-thioFAG (4), the structure
of which was identical to that synthesized by the original
method,24 was obtained from 20 in 34% yield (Scheme
5).
fusion reaction, after evaporation of the solvents, to give
a 1:1 mixture of R- and â-anomers of 16 in 61% yield (data
not shown). The loss of stereoselectivity in the glycosyl-
ation reaction may have been due to the pretreatment
of 15 in refluxing carbon tetrachloride, which caused the
anomerization of R-bromide (vide supra). After many
efforts were made to optimize the reaction conditions, we
found that the glycosylated product 16 was formed in a
yield of 59% when the 1-bromide 15 was fused with the
persilylated N4-acetylcytosine at 80 °C for 5 h under
reduced pressure. Compound 16 was deprotected by
treatment with concentrated NH4OH/MeOH to give the
desired 4′-thioFAC (2) as an anomeric mixture. HPLC
analysis of the crude products clearly showed that the
desired â-anomer of 4′-thioFAC (2) was predominantly
formed (R:â ) 1:4), as we expected. The fusion reaction
was critical for the progression of the glycosylation
reaction and the predominant formation of â-anomer. The
structures of â-2 and R-2 were confirmed by comparison
1
of the instrumental analysis data, e.g., H NMR, FAB
mass spectra, and retention time of HPLC, after isolation
(Scheme 3).
Syn th esis of 4′-Th ioF AG. Recently, we have reported
a novel antiviral 4′-thionucleoside, 4′-thioFAG (4), which
has potent antiviral activities against herpes simplex
virus type 1 (HSV-1) and type 2 (HSV-2) in vitro.23,24 4′-
ThioFAG (4) also has potent activity against human
cytomegalovirus (HCMV).23,24 To further evaluate this
compound, in vivo assays were needed. However, the
original synthesis of 4 had several problems,24 which
needed to be overcome before 4 could be prepared on a
multigram scale. As in the case of 4′-thioFAC, the biggest
problem was the glycosylation step. The Lewis acid-
catalyzed glycosylation reaction of 2,6-diaminopurine
with 2-fluoro-4-thioarabinose derivative 17 gave the
desired 4′-thionucleoside 18 in 39% yield as a mixture of
R- and â-anomers (R:â ) 1:1.5)24 (Scheme 4).
The success of the improved synthesis of 4′-thioFAC
encouraged us to apply this method to the synthesis of
Syn th esis of 4′-Th iocyta r a zid . As a final target of
this project, we chose 4′-thiocytarazid. Cytarazid, (1-(2-
azido-2-deoxy-â-D-arabino-pentofuranosyl)cytosine), a par-
ent compound of 4′-thiocytarazid, is known to have potent
antineoplastic activities against solid tumors as well as
(23) Machida, H.; Ashida, N.; Miura, S.; Endo, M.; Yamada, K.;
Kitano, K.; Yoshimura, Y.; Sakata, S.; Ijichi, O.; Eizuru, Y. Antiviral
Res., 1998, 39, 129-137.
(24) Yoshimura, Y.; Kitano, K.; Yamada, K.; Sakata, S.; Miura, S.;
Ashida, N.; Machida, H. submitted.

7916 J . Org. Chem., Vol. 64, No. 21, 1999
Yoshimura et al.
Sch em e 5
Sch em e 7a
a
(a) H2SO4, Ac2O, AcOH; (b) silylated N4-acetylcytosine, TM-
SOTf, CH3CN, reflux; (c) concntrated NH4OH, MeOH.
Sch em e 6
Ch a r t 2
leukemia.25,26 Thus, we selected 4′-thiocytarazid as a
potential antitumor agent and attempted to synthesize
it in our previous study. However, the introduction of an
azido group at the 2-position of the key intermediate 21
by the Mitsunobu reaction resulted in the unexpected
formation of a ribo-azido derivative 22.9 Consequently,
we obtained 4′-thioazidocytidine, which was less active
against tumor cell lines9 (Scheme 6).
In contrast to the previous synthesis, in the above-
mentioned alternative synthesis, the 2′-substituent is
introduced in an earlier stage. In addition, the original
synthesis of cytarazid by Bobek et al.25 was achieved by
adapting a reaction scheme similar to Watanabe’s FAC
synthesis, where a fluorine is simply substituted for an
azido group. Therefore, we were hopeful that 4′-thio-
cytarazid could be synthesized using the alternative
synthesis.
The introduction of an azido functional group at the
3-position of 6 has already been reported by Bobek et al.25
However, we thought that we might be able to apply the
reaction we used for the fluorination of 6 to C3 azidation,
in which nucleophilic displacement of the 3-imidazoyl
sulfate 7 was simply performed with sodium azide,
instead of potassium fluoride. Indeed, treatment of 7 with
sodium azide in refluxing 2-methoxyethanol gave a
3-azido derivative 23 in 87% yield from 6. As depicted in
Scheme 1, the 3-azido derivative 23 was further con-
verted to 2-azido-4-thioarabinose 28, and the reaction
scheme was identical to that in the synthesis of 13. All
of the chemical yields in each step were comparable to
those in the 2-fluoro series.
acetolysis of 28, was subjected to the glycosylation
reaction. Although we tried the nucleophilic substitution
of the corresponding 1-bromide prepared from 29 with
persilylated acetylcytosine, the reaction gave none of the
desired product. This seemed to be due to the instability
of the generated 1-bromide. Thus, we prepared the
corresponding 1-chloride, the nucleophilic displacement
of which with persilylated cytosine derivative was also
unsuccessful (data not shown). Alternatively, compound
29 was directly coupled with persilylated acetylcytosine
in the presence of TMS triflate to give protected 4′-
thiocytarazid 30. By simple silica gel column chroma-
tography, the anomers of 30 could be separated, giving
â-30 and R-30 in 19% and 39% yields, respectively, from
29. Both â- and R-30 were deblocked by treatment with
concentrated NH4OH/MeOH to give â- and R-anomers of
4′-thiocytarazid 5 in 79% and 93% yields, respectively
(Scheme 7).
The antineoplastic activities of 4′-thiocytarazid 5 were
evaluated, and the results are summarized in Table 1,
in comparison with those of araC 31, cytarazid 32, 4′-
thioazidocytidine 33, and 4′-thioFAC 2 (Chart 2 and
Table 1). The â-anomer of 4′-thiocytarazid â-5 showed
only moderate inhibitory activity against human T-cell
leukemia (CCRF-HSB-2), which was comparable to that
of ribo-isomer 33. Against solid tumor (KB cells), both
As with the alternative synthesis of 4′-thioFAC, a
1-acetoxy-2-azido derivative 29, which was obtained by
(25) (a) Bobek, M.; Martin, V. Tetrahedron Lett. 1978, 1919-1922.
(b) Bobek, M.; Cheng, Y. C.; Bloch, A. J . Med. Chem. 1978, 21, 597-
598.
(26) Matsuda, A.; Yasuoka, J .; Sasaki, T.; Ueda, T. J . Med. Chem.
1991, 34, 999-1002.

Synthesis of 4′-Thionucleosides
J . Org. Chem., Vol. 64, No. 21, 1999 7917
Ta ble 1. An titu m or Activities of 2′-Su bstitu ted
4′-Th iocytid in es
by sodium bicarbonate, the insoluble salts were removed by
filtration and the filtrate was concentrated under reduced
pressure. The residue was extracted with CHCl3 (×2), and the
combined organic layers were washed with brine and then
dried (Na2SO4). After the filtrate was concentrated under
reduced pressure, the residue was passed through a silica gel
column (3.7 × 14 cm, 50-67% AcOEt in n-hexane) to give a
5,6-diol derivative (1.81 g, 90%). To a solution of the 5,6-diol
derivative (1.74 g, 7.83 mmol), pyridine (697 µL, 8.61 mmol),
and DMAP (10 mg) in CH2Cl2 (30 mL) was added dropwise
BzCl (992 µL, 8.61 mmol) in CH2Cl2 (20 mL) at 0 °C. The
mixture was stirred at 0 °C for 5 h. After the reaction was
quenched with MeOH, the mixture was extracted with CHCl3
(×3). The combined organic layers were washed with 0.5 N
HCl (×2), saturated NaHCO3, and brine and dried (Na2SO4).
After concentration, the residue was purified over a silica gel
column (8.1 × 12 cm, 10-20% AcOEt in n-hexane) to give 9
(2.15 g, 91%) as a crystal: mp 130-132 °C (crystallized from
n-hexane); 1H NMR (CDCl3) δ ppm 8.09-8.05 (2H, m), 7.60-
7.42 (3H, m), 5.99 (1H, d, J ) 3.9 Hz), 5.14 (1H, dd, J ) 2.0,
49.8 Hz), 4.74-4.70 (2H, m), 4.46 (1H, dd, J ) 5.9, 12.2 Hz),
4.27-4.18 (2H, m), 2.83 (1H, br, D2O exchangeable), 1.47, 1.33
(each 3H, s); FAB-MS m/z 327 (M + H+). Anal. Calcd for
IC50 (µg/mL)
compd
no.
CCRF-HSB-2
KB
42
>100
82
0.015
â-4′-thiocytarazid
R-4′-thiocytarazid
4′-thioazidocytidine
4′-thioFAC
â-5
R-5
33
2
32
31
7.8
>100
8.6
0.051
0.15
0.052
cytarazid
1.0
0.26
araC
â-5 and 33 were inactive. The R-anomer of 5 showed no
activity against either cell line. Compared with its parent
compound cytarazid 31, 4′-thiocytarazid â-5 was 50 times
less active against CCRF-HSB-2 and 40 times less active
against KB cells. These results suggest that the antineo-
plasic activities of 4′-thiocytidine analogues depend more
closely on the nature of the 2′-substituents than those of
their 4′-oxy counterparts. Deoxycytidine kinase, a key
enzyme that converts deoxycytidine antimetabolites to
their active form, may strictly recognize the stereo- and
electrochemical properties of 2′-substituents of 4′-thio-
cytidines.
C
16H19FO6: C, 58.89; H, 5.87. Found: C, 58.78; H, 5.82.
5,6-An h yd r o-3-d eoxy-3-flu or o-1,2-O-isop r op ylid en e-â-
L-id o-p en tofu r a n ose (10). To a solution of 9 (8.80 g, 27.0
mmol) in pyridine (45 mL) was added MsCl (3.34 mL, 43.2
mmol), and the mixture was stirred at room temperature for
3.5 h. After quenching with ice-water, the solvent was
removed under reduced pressure. The residue was partitioned
between AcOEt and water. The organic layer was washed with
0.5 N HCl (×2), saturated NaHCO3, and brine and then dried
(Na2SO4). After the filtarte was concentrated under reduced
pressure, the residue was dissolved in MeOH (50 mL) contain-
ing NaOMe (28%, 6.6 mL, 32.4 mmol). The mixture was stirred
at room temperature for 40 min. After evaporation, the residue
was partitioned between AcOEt and water, washed with brine,
and then dried (Na2SO4). After the filtrate was concentrated
under reduced pressure, the residue was purified over a silica
gel column (200 mL, 10-20-40-60% AcOEt in n-hexane) to
give 10 (4.84 g, 89%) as a syrup: 1H NMR (CDCl3) δ ppm 6.04
(1H, d, J ) 3.9 Hz), 4.96 (1H, dd, J ) 2.4, 50.3 Hz), 4.71 (1H,
dd, J ) 3.9, 11.2 Hz), 3.89 (1H, ddd, J ) 2.4, 5.9, 30.3 Hz),
3.22 (1H, ddd, J ) 2.9, 4.4, 5.9 Hz), 2.88 (1H, t, J ) 4.4 Hz),
2.71 (1H, dd, J ) 2.9, 4.9 Hz), 1.47, 1.33 (each 3H, s); FAB-
MS m/z 205 (M + H+). Anal. Calcd for C9H13FO4: C, 52.94; H,
6.42. Found: C, 52.67; H, 6.17.
5,6-An h yd r o-3-d eoxy-3-flu or o-1,2-O-isop r op ylid en e-5-
th io-r-D-glu co-p en tofu r a n ose (11). A mixture of 10 (405 mg,
1.98 mmol) and thiourea (151 mg, 1.98 mmol) in MeOH (10
mL) was kept under reflux for 7 h. After cooling to room
temperature, the solvent was removed under reduced pressure.
The residue was partitioned between AcOEt and water. The
organic layer was washed with brine and dried (Na2SO4). After
the filtrate was concentrated under reduced pressure, the
residue was purified over a silica gel column (2.1 × 10 cm, 5%
AcOEt in n-hexane) to give 11 (386 mg, 88%) as a syrup: 1H
NMR (CDCl3) δ ppm 6.01 (1H, d, J ) 3.9 Hz), 4.93 (1H, dd, J
) 2.0, 49.8 Hz), 4.72 (1H, dd, J ) 3.9, 10.3 Hz), 3.60 (1H, ddd,
J ) 2.4, 8.8, 28.3 Hz), 3.10 (1H, dt, J ) 5.4, 8.8 Hz), 2.65 (1H,
dd, J ) 1.5, 5.8 Hz), 2.41 (1H, d, J ) 5.4 Hz), 1.46, 1.32 (each
3H, s); FAB-MS m/z 221 (M + H+). Anal. Calcd for C9H13FO3S‚
0.13H2O: C, 48.56; H, 6.00. Found: C, 48.59; H, 5.76.
In conclusion, we have adapted an alternative syn-
thetic route for 4′-thionucleosides, which is more practical
than our original method. Using this method, we syn-
thesized 4′-thioFAC, 4′-thioFAG, and 4′-thiocytarazid.
Although the antineoplastic activity of 4′-thiocytarazid
was rather disappointing, this compound could be syn-
thesized by the alternative method but not by the original
method.
Exp er im en ta l Section
Gen er a l In for m a tion . Melting points are uncorrected. 1H
NMR spectra were recorded at 400 MHz (1H) and at 100 MHz
(13C) using CDCl3 or DMSO-d6 with tetramethylsilane as
internal standard. Mass spectra were obtained by the fast
atom bombardment (FAB) mode. Silica gel for chromatography
was Merck Kieselgel 60. All anhydrous reactions were per-
formed under an argon atmosphere.
3-Deoxy-3-flu or o-1,2:5,6-d i-O-isop r op ylid en e-r-D-glu co-
p en tofu r a n ose (8). To a solution of 1,2:5,6-di-O-isopropyl-
idene-R-D-allo-pentofuranose (6) (19.48 g, 74.8 mmol) in CH2Cl2
(200 mL) was added sulfuryl chloride (12.16 mL, 151 mmol).
After 30 min of stirring at 0 °C, imidazole (50.92 g, 748 mmol)
was added. The whole was stirred at room temperature for
2.5 h. The reaction was quenched with saturated NaHCO3 and
extracted with CH2Cl2 (×3). The combined organic layers were
washed with brine and dried (Na2SO4). After being concen-
trated under reduced pressure, the residue was dissolved in
2-methoxyethanol (200 mL). To this mixture was added
potassium fluoride (spray-dried, 43.44 g, 748 mmol). The
mixture was kept under reflux for 7 h. After cooling to room
temperature, the solvent was removed under reduced pressure.
The residue was partitioned between AcOEt and water, and
the organic layer was washed with brine and then dried (Na2-
SO4). After evaporation of the solvents, the residue was
purified over a silica gel column (8.0 × 9.5 cm, 4% AcOEt in
n-hexane) to give 8 (15.05 g, 77%) as a syrup: 1H NMR (CDCl3)
δ ppm 5.95 (1H, d, J ) 3.9 Hz), 5.01 (1H, dd, J ) 2.2, 49.8
Hz), 4.70 (1H, dd, J ) 3.9, 10.7 Hz), 4.29 (1H, ddd, J ) 8.3,
5.9, 4.9 Hz), 4.12 (1H, dd, J ) 5.9, 8.8 Hz), 4.11 (1H, ddd, J )
2.2, 8.3, 29.0 Hz), 4.03 (1H, dd, J ) 4.9, 8.8 Hz), 1.50, 1.45,
1.37, 1.33 (each 3H, s); FAB-MS m/z 263 (M + H+). Anal. Calcd
for C12H19FO5‚0.25H2O: C, 54.03; H, 7.37. Found: C, 54.31;
H, 6.97.
3-Deoxy-5,6-di-O,S-acetyl-3-flu or o-1,2-O-isopr opyliden e-
5-th io-r-D-glu co-p en tofu r a n ose (12). A mixture of 11 (7.12
g, 22.1 mmol) and KOAc (3.46 g, 35.3 mmol) in acetic
anhydride (75 mL) containing acetic acid (15 mL) was kept at
120 °C for 20 h. After cooling to room temperature, the solvent
was removed under reduced pressure. The residue was par-
titioned between AcOEt and water. The organic layer was
washed with saturated NaHCO3 and brine and dried (Na2SO4).
After the filtrate was concentrated under reduced pressure,
the residue was purified over a silica gel column (220 mL, 15-
20% AcOEt in n-hexane) to give 12 (5.20 g, 73%) as a crystal:
6-O-Ben zoyl-3-d eoxy-3-flu or o-1,2-O-isop r op ylid en e-r-
D-glu co-p en tofu r a n ose (9). A mixture of 8 (2.38 g, 9.07
mmol) in THF (10 mL) and 2 M HCl (10 mL) was stirred at
room temperature for 1 h. After the mixture was neutralized
1
mp 89-90 °C (crystallized from AcOEt-n-hexane); H NMR

7918 J . Org. Chem., Vol. 64, No. 21, 1999
Yoshimura et al.
(CDCl3) δ ppm 5.98 (1H, d, J ) 3.9 Hz), 4.96 (1H, dd, J ) 2.0,
49.6 Hz), 4.68 (1H, dd, J ) 3.9 Hz), 4.46-4.37 (3H, m), 4.11
(1H, dt, J ) 3.9, 4.8 Hz), 2.36 (3H, s), 2.06 (3H, s), 1.49, 1.33
(each 3H, s); FAB-MS m/ z 323 (M + H+). Anal. Calcd for
C13H19O6SF: C, 48.44; H, 5.94. Found: C, 48.49; H, 6.00.
Meth yl 2-d eoxy-3,5-d i-O-ben zoyl-2-flu or o-4-th io-r- a n d
â-D-a r a bin o-p en tofu r a n osid e (13). A solution of 12 (480 mg,
1.49 mmol) in aqueous trifluoroacetic acid (90%, 4 mL) was
stirred at 0 °C for 4.5 h. The solvent was removed under
reduced pressure below 20 °C. The residue was partitioned
between AcOEt and water and washed with saturated NaH-
CO3. The water layer was extracted with AcOEt (×2) and the
combined organic layers were washed with saturated NaHCO3
and brine and then dried (Na2SO4). After the filtrate was
concentrated under reduced pressure, the residue was dis-
solved in MeOH (6 mL). To this mixture was added dropwise
NaIO4 (319 mg, 1.49 mmol) in water (6 mL). After 25 min of
stirring at room temperature, glycerin (0.2 mL) was added.
The whole was stirred at room temperature for 25 min. After
MeOH (30 mL) was added, insoluble salts were removed by
suction. Most of MeOH was removed under reduced pressure,
and the residual water solution was extracted with CHCl3 (×3).
The combined organic layer was washed with brine, dried
(Na2SO4), and concentrated. The residue was dissolved in
methanolic HCl (prepared from 0.35 mL of acetyl chloride and
6.65 mL of MeOH), and the mixture was kept under reflux
for 2 h. After cooling to room temperature, the mixture was
neutralized with NaHCO3. The insoluble salts were removed
by filtration. After the filtrate was concentrated under reduced
pressure, the residue was coevaporated with pyridine (×3) and
dissolved in pyridine (7 mL). To this mixture was added
benzoyl chloride (3.61 mL, 31.3 mmol). The mixture was stirred
at room temperature for 1 h. MeOH was added to the mixture,
and the whole was stirred at room temperature for 30 min.
After concentration under reduced pressure, the residue was
partitioned between AcOEt and water. The organic layer was
washed with 0.5 N HCl (×2), saturated NaHCO3, and brine
and dried (Na2SO4). After evaporation, the residue was purified
over a silica gel column (2.4 × 12.5 cm, 2-4% AcOEt in
n-hexane) to give R-13 (less polar, 180 mg) and â-13 (more
polar, 231 mg, total 71%).
7.8 Hz), 3.74 (0.58H, q, J ) 6.4 Hz), 2.12, 2.11 (total 3H, s);
FAB-MS m/ z 419 (M + H+). Anal. Calcd for C21H19O6SF: C,
60.28; H, 4.58. Found: C, 60.06; H, 4.56.
4-Acet yl-1-(2-d eoxy-3,5-d i-O-b en zoyl-2-flu or o-4-t h io-
r,â-D-a r a bin o-p en tofu r a n osyl)cytosin e (16). To a solution
of 14 (3.00 g, 7.20 mmol) in CH2Cl2 (20 mL) was added 30%
HBr/acetic acid solution (6.0 mL) at room temperature. The
mixture was stirred at room temperature for 20 min. The
reaction was quenched by the addition of ice water. The whole
was extracted with CH2Cl2 twice, and the combined organic
layers were washed with saturated NaHCO3 and ice water and
then dried (Na2SO4). The solvent was removed under reduced
pressure and the residual 15 was dissolved in CH2Cl2. To a
solution of silylated N4-acetylcytosine (prepared from N4-
acetylcytosine (1.65 g, 10.8 mmol) by refluxing with bis-
(trimethylsilyl)acetamide (5.3 mL, 21.5 mmol) in 1,2-dichlo-
roethane (20 mL) for 3 h) was added the CH2Cl2 solution of
15. After evaporation, the neat mixture was kept at 80 °C for
5 h under reduced pressure (less than 4 mmHg). After cooling
to room temperature, CHCl3 was added, and the insoluble
materials were removed by filtration. The filtrate was con-
centrated, and the residue was purified over a silica gel column
(200 mL, 0-5% MeOH in CHCl3) to give 16 (2.17 g, 59%) as
an amorphous foam: 1H NMR (CDCl3) δ ppm 8.58 (0.2 H, br,
D2O exchangeable), 8.39 (0.8H, br, D2O exchangeable), 8.38
(0.8H, d, J ) 7.3 Hz), 8.37 (0.2H, d, J ) 6.8 Hz), 8.12-8.04
(4H, m), 7.83-7.38 (7H, m), 6.96 (0.8H, dd, J ) 3.9, 23.4 Hz),
6.38 (0.2H, dd, J ) 2.0, 14.2 Hz), 5.92 (0.8H, dt, J ) 2.0, 9.3
Hz), 5.86 (0.2H, dt, J ) 2.9, 12.2 Hz), 5.52 (0.2H, dt, J ) 2.0,
46.4 Hz), 5.39 (0.8H, ddd, J ) 2.4, 3.4, 49.3 Hz), 4.75-4.56
(2H, m), 4.35-4.32 (0.2H, m), 4.04 (0.8H, t, J ) 7.8 Hz), 2.26,
2.25 (total 3H, s); FAB-MS m/z 511 (M + H+). Anal. Calcd for
C
25H22N3O6SF‚1.25H2O: C, 56.23; H, 4.62; N, 7.87. Found: C,
55.92; H, 4.27; N, 8.16.
1-(2-Deoxy-2-flu or o-4-th io-r- a n d â-D-a r a bin o-p en to-
fu r a n osyl)cytosin e (r- a n d â-4′-Th ioF AC, r- a n d â-2). A
suspension of 16 (2.17 g, 4.24 mmol) in MeOH (30 mL) and
concentrated NH4OH (30 mL) was stirred at room temperature
for 2 days. The solvents were removed under reduced pressure,
and the residue was coevaporated with EtOH (×3). The residue
was purified over a silica gel column (200 mL, 10-33% MeOH
in CHCl3). An anomeric mixture of 2 was separated by an ODS
column (Wako sil 40C18, Wako Pure Chemical Industries, Ltd.,
J apan; 400 mL, H2O only). The unseparable fractions were
collected. The ODS column chromatography was repeated to
give â-2 (613 mg, 55%) and R-2 (152 mg, 14%). All of the
instrumental analyses data for â-2 and R-2 were identical to
those reported previously.
Da ta for r-13: an amorphous foam; 1H NMR (CDCl3) δ
ppm 8.04-7.95 (4H, m), 7.60-7.33 (6H, m), 5.80-5.73 (1H,
m), 5.28 (1H, ddd, J ) 2.4, 3.9, 48.3 Hz), 5.22 (1H, dd, J ) 2.0,
13.7 Hz), 4.59 (1H, dd, J ) 6.8, 11.2 Hz), 4.49 (1H, ddd, J )
1.0, 6.4, 11.2 Hz), 4.05 (1H, q, J ) 6.8 Hz), 3.42 (3H, s); FAB-
MS m/ z 391 (M + H+). Anal. Calcd for C20H19O5SF: C, 61.53;
H, 4.91. Found: C, 61.29; H, 4.87.
Da ta for â-13: mp 75-80 °C (crystallized from AcOEt-n-
hexane); 1H NMR (CDCl3) δ ppm 8.04-7.97 (4H, m), 7.60-
7.31 (6H, m), 6.09 (1H, ddd J ) 6.4, 8.8, 12.2 Hz), 5.29 (1H,
ddd, J ) 4.4, 8.8, 51.8 Hz), 4.96 (1H, d, J ) 4.4 Hz), 4.64 (1H,
dd, J ) 6.4, 11.2 Hz), 4.50 (1H, dd, J ) 6.4, 11.2 Hz), 3.69
(1H, q, J ) 6.4 Hz), 3.44 (3H, s); FAB-MS m/ z 391 (M + H+).
Anal. Calcd for C20H19O5SF: C, 61.53; H, 4.91. Found: C,
61.94; H, 4.93.
1-O-Acet yl-2-d eoxy-3,5-d i-O-b en zoyl-2-flu or o-4-t h io-
r,â-D-a r a bin o-p en tofu r a n ose (14). A mixture of 13 (1.09 g,
2.80 mmol), acetic anhydride (10 mL), acetic acid (10 mL), and
sulfuric acid (0.3 mL) was stirred at room temperature for 1
h. After being neutralized with NaOAc, the mixture was
stirred at room temperature for 1 h and was partitioned
between CH2Cl2 and water. The water layer was extracted
with CH2Cl2 (×3). The combined organic layers were washed
with saturated NaHCO3 (×3) and brine and dried (Na2SO4).
After concentration, the residue was coevaporated with MeOH
(×2) and CH2Cl2. Crystallization of the residue from n-hexane
gave 14 (933 mg, 80%): mp 85-93 °C; 1H NMR (CDCl3) δ ppm
8.06-7.94 (4H, m), 7.62-7.30 (6H, m), 6.24 (0.42H, dd, J )
2.0, 14.2 Hz), 6.18 (0.58H, d, J ) 4.4 Hz), 6.08 (0.58H, ddd, J
) 7.3, 9.3, 11.7 Hz), 5.85 (0.42H, dt, J ) 3.9, 12.2 Hz), 5.39
(0.42H, ddd, J ) 2.0, 3.9, 47.9 Hz), 5.31 (0.58H, ddd, J ) 4.4,
9.3, 50.8 Hz), 4.69 (0.58H, dd, J ) 6.4, 11.2 Hz), 4.55 (0.42H,
dd, J ) 7.8, 11.7 Hz), 4.49 (0.58H, dd, J ) 6.4, 11.2 Hz), 4.47
(0.42H, dd, J ) 1.5, 11.7 Hz), 4.11 (0.42H, ddd, J ) 1.5, 4.4,
9-(2-Deoxy-3,5-d i-O-ben zoyl-2-flu or o-4-th io-r,â-D-a r a -
bin o-p en tofu r a n osyl)-2,6-d ia m in op u r in e (19). Compound
15 was prepared from 14 (6.81 g, 16.3 mmol) as described in
the synthesis of 16. A mixture of 15 (as 16.3 mmol), 2,6-
diaminopurine (7.34 g, 48.9 mmol), TMS triflate (18.9 mL, 97.6
mmol), and 4A molecular sieves (18.7 g) in CH3CN (150 mL)
was kept at 60 °C overnight. After cooling to room tempera-
ture, the reaction was quenched with saturated NaHCO3. The
whole was extracted with CHCl3 (×3), and the organic layer
was washed with brine and then dried (Na2SO4). After
concentration under reduced pressure, the residue was purified
over a silica gel column (400 mL, 0-1-3% MeOH in CHCl3)
to give 19 (6.20 g, 75%) as an amorphous foam: 1H NMR
(CDCl3) δ ppm 8.09-7.39 (11H, m), 6.52 (0.5H, dd, J ) 3.9,
21.0 Hz), 6.19 (0.5H, dd, J ) 2.9, 14.2 Hz), 6.14 (0.5H, dt, J )
2.9, 9.3 Hz), 5.91 (0.5H, dd, J ) 3.9, 6.3 Hz), 5.87 (0.5H, dt, J
) 3.9, 63.5 Hz), 5.37 (2H, br, D2O exchangeable), 5.31 (0.5H,
dt, J ) 3.9, 49.3 Hz), 4.77 (2H, br, D2O exchangeable), 4.84-
4.55 (2H, m), 4.41-4.36 (0.5H, m), 4.05-4.01 (0.5H, m); FAB-
MS m/z 509 (M + H+). Anal. Calcd for C24H21N6O4SF‚
1.25H2O: C, 54.28; H, 4.46; N, 15.83. Found: C, 54.62; H, 4.22;
N, 15.46.
9-(2-Deoxy-2-flu or o-4-th io-r,â-D-a r a bin o-p en tofu r a n o-
syl)-2,6-d ia m in op u r in e (20). A suspension of 19 (6.20 g, 12.2
mmol) in MeOH (160 mL) and concentrated NH4OH (160 mL)
was stirred at room temperature overnight. After MeOH was
evaporated under reduced pressure, the mixture was washed

Synthesis of 4′-Thionucleosides
J . Org. Chem., Vol. 64, No. 21, 1999 7919
with CHCl3 (×3). Insoluble materials were removed by filtra-
tion. The solvents were removed under reduced pressure.
Crystallization of the residue from H2O gave 20 (3.02 g, 82%)
as an anomeric mixture: 1H NMR (DMSO-d3 + D2O) δ ppm
8.18 (0.5H, s), 8.10 (0.5H, s), 6.06 (0.5H, t, J ) 6.8 Hz), 6.00
(0.5H, dd, J ) 6.3, 16.6 Hz), 5.52 (0.5H, dt, J ) 6.8, 52.2 Hz),
5.09 (0.5H, ddd, J ) 5.9, 7.3, 50.8 Hz), 4.42 (0.5H, dt, J ) 7.3,
12.2 Hz), 4.19 (0.5H, dt, J ) 7.3, 13.7 Hz), 3.86-3.73 (2H, m),
3.52 (0.5H, dd, J ) 7.3, 11.2 Hz), 3.29 (0.5H, q, J ) 5.8 Hz);
FAB-MS m/z 301 (M + H+). Anal. Calcd for C10H13N6O3FS‚
0.5H2O: C, 38.83; H, 4.56; N, 27.17. Found: C, 38.64; H, 4.64;
N, 27.21.
9-(2-Deoxy-2-flu or o-4-th io-â-D-a r a bin o-pen tofu r an osyl)-
gu a n in e (4). To a suspension of 20 (3.02 g, 10.1 mmol) in
Tris-HCl buffer (pH 7.0, 420 mL) was added adenosine
deaminase (from bovine spleen, EC 3.5.4.4, purchased from
Sigma Co., Ltd., 6.0 mL, 1400 units). The mixture was stirred
at room temperature for 4 h. Another Tris-HCl buffer (30 mL)
was added, and the mixture was stirred for 2 h. The reaction
mixture was applied to a top of ODS column (Chromatorex,
Fuji Silysia Chemical Ltd., 500 mL). From the eluate contain-
ing 1% CH3CN, compound 4 (710 mg) was obtained after
evaporation of the solvents. Similary, from the eluate contain-
ing 3% CH3CN, unreacted 20 (1.46 g) was obtained. Hydrolysis
of the recovered 20 by adenosine deaminase was repeated to
give 4 (330 mg, total 34%) and R-20 (0.98 g, 32%). All of the
instrumental analyses data for 4 and R-20 were identical to
those reported previously.
3-Azid o-3-d eoxy-1,2:5,6-d i-O-isop r op ylid en e-r-D-glu co-
p en tofu r a n ose (23). To a solution of 1,2:5,6-di-O-isopropyl-
idene-R-D-allo-pentofuranose (6) (10.4 g, 40 mmol) in CH2Cl2
(100 mL) was added sulfuryl chloride (6.48 mL, 80 mmol).
After 30 min of stirring at 0 °C, imidazole (27.2 g, 400 mmol)
was added. The whole was stirred at room temperature
overnight. The reaction was quenched with saturated NaHCO3
and extracted with CH2Cl2 (×3). The combined organic layers
were washed with brine and dried (Na2SO4). After concentra-
tion under reduced pressure, the residue was dissolved in
2-methoxyethanol (100 mL). To this mixture was added sodium
azide (7.80 g, 120 mmol). The mixture was kept under reflux
for 7 h. After cooling to room temperature, the solvent was
removed under reduced pressure. The residue was partitioned
between AcOEt and water, and the organic layer was washed
with brine and then dried (Na2SO4). After evaporation, the
residue was purified over a silica gel column (7.3 × 10.5 cm,
4% AcOEt in n-hexane) to give 23 (9.90 g, 87%) as a syrup:
1H NMR (CDCl3) δ ppm 5.86 (1H, d, J ) 3.4 Hz), 4.62 (1H, d,
J ) 3.9 Hz), 4.24 (1H, ddt, J ) 1.0, 4.9, 5.9 Hz), 4.14 (1H, dd,
J ) 5.9, 8.8 Hz), 4.11-4.08 (2H, m), 3.98 (1H, dd, J ) 4.9, 8.8
Hz), 1.51, 1.44, 1.37, 1.33 (each 3H, s); FAB-MS m/z 286 (M +
H+); IR (neat) 2112 cm-1 (N3). Anal. Calcd for C12H19N3O5‚
0.2H2O: C, 49.89; H, 6.77; N, 14.54. Found: C, 50.05; H, 6.55;
N, 14.42.
dd, J ) 3.4, 5.4 Hz), 3.21 (1H, ddd, J ) 2.4, 4.4, 5.4 Hz), 2.89
(1H, t, J ) 4.9 Hz), 2.70 (1H, dd, J ) 2.4, 4.9 Hz), 1.49, 1.33
(each 3H, s); FAB-MS m/z 228 (M + H+); IR (neat) 2114 cm-1
(N3). Anal. Calcd for C9H13N3O4: C, 47.57; H, 5.77; N, 18.49.
Found: C, 47.24; H, 5.53; N, 18.50.
5,6-An h yd r o-3-a zid o-3-d eoxy-1,2-O-isop r op ylid en e-5-
th io-r-D-glu co-p en tofu r a n ose (26). Compound 26 was syn-
thesized from 25 (3.79 g, 16.7 mmol) as described in the
synthesis of 11. The residue was purified over a silica gel
column (4.7 × 11.5 cm, 5% AcOEt in n-hexane) to give 26 (3.52
g, 87%) as a syrup: 1H NMR (CDCl3) δ ppm 5.93 (1H, d, J )
3.4 Hz), 4.66 (1H, d, J ) 3.9 Hz), 4.02 (1H, d, J ) 3.4 Hz),
3.62 (1H, dd, J ) 3.4, 8.8 Hz), 3.10 (1H, ddd, J ) 4.9, 5.9, 8.8
Hz), 2.66 (1H, dd, J ) 1.5, 6.3 Hz), 2.41 (1H, dd, J ) 1.5, 4.9
Hz), 1.47, 1.32 (each 3H, s); FAB-MS m/z 244 (M + H+); IR
(neat) 2110 cm-1 (N3). Anal. Calcd for C9H13N3O3S‚0.25H2O:
C, 43.63; H, 5.49; N, 16.96. Found: C, 43.33; H, 5.26; N, 16.89.
3-Azido-3-deoxy-5,6-di-O,S-acetyl-1,2-O-isopr opyliden e-
5-th io-r-D-glu co-p en tofu r a n ose (27). Compound 27 was
synthesized from 26 (3.329 g, 13.6 mmol) as described in the
synthesis of 12. The residue was purified over a silica gel
column (4.3 × 12 cm, 5-10-15% AcOEt in n-hexane) to give
27 (3.31 g, 70%); mp 94-95 °C (crystallized from AcOEt-n-
hexane); 1H NMR (CDCl3) δ ppm 5.89 (1H, d, J ) 3.4 Hz),
4.66 (1H, d, J ) 3.9 Hz), 4.43 (1H, dd, J ) 3.4, 11.7 Hz), 4.39
(1H, dd, J ) 3.4, 10.7 Hz), 4.33 (1H, dd, J ) 4.9, 11.7 Hz),
4.06 (1H, ddd, J ) 3.4, 4.9, 10.7 Hz), 4.00 (1H, d, J ) 2.9 Hz),
2.38, 2.05 (each 3H, s), 1.50, 1.33 (each 3H, s); FAB-MS m/z
346 (M + H+); IR (KBr) 2116 cm-1 (N3). Anal. Calcd for
C
13H19N3O6S: C, 45.21; H, 5.54; N, 12.17. Found: C, 45.29;
H, 5.54; N, 12.07.
Met h yl 2-Azid o-2-d eoxy-4-t h io-r,â-D-a r a bin o-p en t o-
fu r a n osid e (28). Compound 28 was synthesized from 27 (3.09
g, 8.95 mmol) as described in the synthesis of 13. The residue
was purified over a silica gel column (4.3 × 16 cm, 5-10%
AcOEt in n-hexane) to give R-28 (757 mg) and â-28 (1.25 g,
total 54%).
Da ta for r-28: mp 43-47 °C (crystallized from AcOEt-n-
hexane); 1H NMR (CDCl3) δ ppm 8.08-7.93 (4H, m), 7.61-
7.28 (6H, m), 5.52 (1H, t, J ) 7.3 Hz), 5.06 (1H, d, J ) 4.9
Hz), 4.60 (1H, dd, J ) 6.4, 11.7 Hz), 4.44 (1H, dd, J ) 6.4,
11.7 Hz), 4.34 (1H, dd, J ) 4.9, 7.3 Hz), 4.06 (1H, dt, J ) 6.4,
7.8 Hz), 3.42 (3H, s); FAB-MS m/z 414 (M + H+); IR (KBr)
2112 cm-1 (N3). Anal. Calcd for C20H19N3O5S‚0.25H2O: C,
57.47; H, 4.70; N, 10.05. Found: C, 57.40; H, 4.71; N, 9.83.
Da ta for â-28: mp 133-136 °C (crystallized from AcOEt-
n-hexane); 1H NMR (CDCl3) δ ppm 8.03-7.92 (4H, m), 7.59-
7.27 (6H, m), 6.04 (1H, dd, J ) 6.8, 10.3 Hz), 4.95 (1H, d, J )
3.9 Hz), 4.60 (1H, dd, J ) 6.4, 11.2 Hz), 4.51 (1H, dd, J ) 6.4,
11.7 Hz), 4.08 (1H, dd, J ) 3.9, 9.8 Hz), 3.68 (1H, q, J ) 6.4
Hz), 3.42 (3H, s); FAB-MS m/z 414 (M + H+); IR (KBr) 2104
cm-1 (N3). Anal. Calcd for C20H19N3O5S: C, 58.10; H, 4.63; N,
10.16. Found: C, 58.20; H, 4.59; N, 10.09.
3-Azid o-6-O-ben zoyl-3-d eoxy-1,2-O-isop r op ylid en e-r-D-
glu co-p en tofu r a n ose (24). Compound 24 was synthesized
from 23 (9.80 g, 34.3 mmol) as described in the synthesis of 9.
The residue was purified over a silica gel column (8.1 × 12
cm, 10-20% AcOEt in n-hexane) to give 24 (9.55 g, 80%) as a
syrup: 1H NMR (CDCl3) δ ppm 8.07-8.05 (2H, m), 7.60-7.42
(3H, m), 5.90 (1H, d, J ) 3.4 Hz), 4.72 (1H, dd, J ) 2.0, 11.7
Hz), 4.65 (1H, d, J ) 3.9 Hz), 4.45 (1H, dd, J ) 5.4, 11.7 Hz),
4.26-4.17 (3H, m), 2.90 (1H, d, J ) 4.4 Hz, D2O exchangeable),
1.49, 1.33 (each 3H, s); FAB-MS m/z 350 (M + H+); IR (KBr)
2112 cm-1 (N3). Anal. Calcd for C16H19N3O6‚0.5H2O: C, 53.63;
H, 5.63; N, 11.73. Found: C, 53.96; H, 5.48; N, 11.60.
5,6-An h yd r o-3-a zid o-3-d eoxy-1,2-O-isop r op ylid en e-â-L-
id o-p en tofu r a n ose (25). Compound 25 was synthesized from
24 (8.09 g, 23.2 mmol) as described in the synthesis of 10. The
residue was purified over a silica gel column (4.8 × 15 cm,
10-20-40-60% AcOEt in n-hexane) to give 25 (3.48 g) and a
5-mesyl derivative (1.27 g). Treatment of the 5-mesyl deriva-
tive (1.27 g, 3.80 mmol) with NaH (60%, 152 mg, 3.80 mmol)
in THF (15 mL) for 30 min at 0 °C gave 25 (584 mg, total 77%)
as a syrup: 1H NMR (CDCl3) δ ppm 5.96 (1H, d, J ) 3.4 Hz),
4.67 (1H, d, J ) 3.4 Hz), 3.98 (1H, d, J ) 3.4 Hz), 3.94 (1H,
1-O-Acetyl-2-a zid o-2-d eoxy-3,5-d i-O-ben zoyl-4-th io-r,â-
D-a r a bin o-p en tofu r a n ose (29). Compound 29 was synthe-
sized from 28 (1.72 g, 4.16 mmol) as described in the synthesis
of 14. The residue was purified over a silica gel column (3.7 ×
12 cm, 10-20% AcOEt in n-hexane) to give 29 (1.63 g, 89%)
as a crystalline solid: mp 136-137 °C (crystallized from
AcOEt-n-hexane); 1H NMR (CDCl3) δ ppm 8.05-7.90 (4H, m),
7.61-7.27 (6H, m), 6.14 (0.7H, d, J ) 4.4 Hz), 6.01-5.97 (1H,
m), 5.60 (0.3H, t, J ) 6.4 Hz), 4.67 (0.7H, dd, J ) 6.4, 11.2
Hz), 4.62 (0.3H, dd, J ) 6.8, 11.7 Hz), 4.52-4.48 (1H, m), 4.43
(0.3H, dd, J ) 6.8, 11.7 Hz), 4.20 (0.7H, dd, J ) 4.4, 10.3 Hz),
4.10 (0.3H, q, J ) 6.8 Hz), 3.33 (0.7H, q, J ) 6.4 Hz), 2.13,
2.12 (total 3H, s); FAB-MS m/z 442 (M + H+); IR (KBr) 2108
cm-1 (N3). Anal. Calcd for C21H19N3O6S: C, 57.14; H, 4.34; N,
9.52. Found: C, 57.38; H, 4.40; N, 9.25.
4-Acet yl-1-(2-a zid o-2-d eoxy-3,5-d i-O-b en zoyl-4-t h io-r
a n d â-D-a r a bin o-p en tofu r a n osyl)cytosin e (30). To a solu-
tion of N4-acetylcytosine (447 mg, 2.92 mmol) in CH3CN (10
mL) was added BSA (1.59 mL, 6.42 mmol), and the mixture
was kept under reflux for 3.5 h. After cooling to room
temperature, 29 (645 mg, 1.46 mmol) and TMSOTf (282 µL,
1.46 mmol) were added. The mixture was kept under reflux

7920 J . Org. Chem., Vol. 64, No. 21, 1999
Yoshimura et al.
for 3 h. After cooling to room temperature, the solvents were
removed under reduced pressure. The residue was dissolved
in CHCl3. To this mixture was added saturated NaHCO3, and
the insoluble materials were removed by suction. The sepa-
rated water layer was extracted with CHCl3 (×3), which was
dried over Na2SO4. After concentration, the residue was
purified over a silica gel column (2.4 × 17.5 cm, 50-66% AcOEt
in n-hexane) to give R-30 (less polar, 305 mg) and â-30 (more
polar, 148 mg, total 58%).
182-185 °C (crystallized from CH3CN-H2O); 1H NMR (DMSO-
d3) δ ppm 8.03 (1H, d, J ) 7.3 Hz), 7.17, 7.09 (total 2H, br,
D2O exchangeable), 6.28 (1H, d, J ) 6.8 Hz), 5.85 (1H, d, J )
5.9 Hz, D2O exchangeable), 5.74 (1H, d, J ) 7.3 Hz), 5.18 (1H,
d, J ) 4.9, 5.4 Hz, D2O exchangeable), 4.23 (1H, dd, J ) 6.8,
9.8 Hz), 3.92 (1H, ddd, J ) 6.3, 8.3, 9.8 Hz), 3.79 (1H, ddd, J
) 3.4, 4.4, 11.7 Hz), 3.68 (1H, dt, J ) 5.9, 11.7 Hz), 3.18 (1H,
ddd, J ) 3.4, 5.9, 8.3 Hz); NOE, irradiate H-6 (8.03 ppm),
observe H-3′ (3.92 ppm; 10.8%); irradiate H-2′ (4.23 ppm),
observe H-4′ (3.18 ppm, 5.3%); irradiate H-3′, observe H-6
(13.6%); irradiate H-4′, observe H-2′ (6.6%); FAB-MS m/z 285
(M + H+); IR (KBr) 2130 cm-1 (N3). Anal. Calcd for C9H12N6O3S‚
0.25H2O: C, 37.43; H, 4.36; N, 29.10. Found: C, 37.20; H, 4.02;
N, 29.13.
1-(2-Azido-2-deoxy-4-th io-r-D-a r a bin o-pen tofu r an osyl)-
cytosin e (r-5). Compound R-30 (193 mg, 0.361 mmol) was
treated as described above to give R-5 (96 mg, 93%) as an
amorphous foam: 1H NMR (DMSO-d3) δ ppm 7.97 (1H, d, J
) 7.3 Hz), 7.27 (2H, br d, D2O exchangeable), 6.07 (1H, d, J )
5.9 Hz, D2O exchangeable), 5.92 (1H, d, J ) 8.8 Hz), 5.81 (1H,
d, J ) 7.8 Hz), 4.96 (1H, d, J ) 4.9, 5.9 Hz, D2O exchangeable),
4.26 (1H, t, J ) 9.3 Hz), 3.87-3.77 (2H, m), 3.59 (1H, dt, J )
3.4, 8.3 Hz), 3.41 (1H, ddd, J ) 6.4, 7.8, 10.3 Hz); FAB-MS
m/z 285 (M + H+); IR (KBr) 2114 cm-1 (N3). Anal. Calcd for
C9H12N6O3S‚0.3(CH3)2CO: C, 38.19; H, 4.47; N, 27.00.
Found: C, 37.89; H, 4.57; N, 26.90.
Da ta for r-30: an amorphous foam; 1H NMR (CDCl3) δ
ppm 8.94 (1H, br, D2O exchangeable), 8.37 (1H, d, J ) 7.8 Hz),
8.07-8.04 (2H, m), 7.83-7.81 (2H, m), 7.59-7.38 (7H, m), 6.16
(1H, d, J ) 3.4 Hz), 5.64 (1H, t, J ) 3.4 Hz), 4.73-4.68 (2H,
m), 4.58 (1H, dd, J ) 7.3, 11.2 Hz), 4.28 (1H, dt, J ) 3.4, 7.3
Hz), 2.27 (3H, s); FAB-MS m/z 535 (M + H+); IR (KBr) 2116
cm-1 (N3). Anal. Calcd for C25H22N6O6S‚0.33AcOEt: C, 56.10;
H, 4.41; N, 14.91. Found: C, 55.86; H, 4.10; N, 14.66.
Da ta for â-30: mp 206-208 °C (crystallized from AcOEt-
1
n-hexane); H NMR (CDCl3) δ ppm 8.43 (1H, d, J ) 7.8 Hz),
8.27 (1H, br, D2O exchangeable), 8.09-8.03 (4H, m), 7.65-
7.46 (6H, m), 7.40 (1H, d, J ) 7.8 Hz), 6.80 (1H, d, J ) 5.9
Hz), 5.77 (1H, t, J ) 4.9 Hz), 4.79 (1H, dd, J ) 6.8, 11.7 Hz),
4.73 (1H, t, J ) 5.4 Hz), 4.65 (1H, dd, J ) 7.3, 11.2 Hz), 3.96
(1H, dt, J ) 4.9, 6.4 Hz), 2.24 (3H, s); FAB-MS m/z 535 (M +
H+); IR (KBr) 2120 cm-1 (N3). Anal. Calcd for C25H22N6O6S‚
0.5H2O: C, 55.24; H, 4.26; N, 15.46. Found: C, 55.45; H, 3.96;
N, 15.14.
Ack n ow led gm en t. The authors are grateful to Prof.
A. Matsuda of Hokkaido University for his useful
suggestions during this work. The authors also acknowl-
edge Dr. K. Kodama of Yamasa Corp. and Prof. H.
Tanaka of Showa University for their encouragement.
1-(2-Azido-2-deoxy-4-th io-â-D-a r a bin o-pen tofu r an osyl)-
cytosin e (4′-Th iocyta r a zid , â-5). A suspension of â-30 (173
mg, 0.324 mmol) in MeOH (8 mL) and concentrated NH4OH
(8 mL) was stirred at room temperature for 2 days. After the
solvents were removed under reduced pressure, the residue
was dissolved in water (50 mL) and purified over an ODS
column (2.2 × 15 cm, H2O only) to give â-5 (73 mg, 79%): mp
J O990958G