A highly efficient synthesis of L-β-2′-deoxy-4′-thio-1′-purine nucleosides as potential antiviral agents

Article abstract of DOI:10.1039/b001240h

L-β-2′-Deoxy-4′-thio-1′-purine nucleosides were synthesized efficiently utilizing the neighboring group effect of the 2-benzoyl-4-thiosugar acetate. The Royal Society of Chemistry 2000.

Full text of DOI:10.1039/b001240h

View Article Online / Journal Homepage / Table of Contents for this issue
A highly efficient synthesis of L-ꢀ-2Ј-deoxy-4Ј-thio-1Ј-purine
nucleosides as potential antiviral agents
Hea Ok Kim,a Lak Shin Jeong,*b Sun Nan Lee,b Soo Jeong Yoo,b Hyung Ryong Moon,b
Kil Soo Kimb and Moon Woo Chunc
1
a College of Medicine, Yonsei University, Seoul 120-752, Korea
b Laboratory of Medicinal Chemistry, College of Pharmacy, Ewha Womans University,
Seoul 120-750, Korea
c College of Pharmacy, Seoul National University, Seoul 151-742, Korea
Received (in Cambridge, UK) 15th February 2000, Accepted 28th March 2000
L-ꢀ-2Ј-Deoxy-4Ј-thio-1Ј-purine nucleosides were synthe-
sized efficiently utilizing the neighboring group effect of
the 2-benzoyl-4-thiosugar acetate.
-4Ј-Thionucleosides in which the furanose oxygen is substi-
tuted by a sulfur atom exhibit interesting biological properties
such as antibiotic,1 antiviral2 and antitumor3 activities, and
also have inherent advantages such as a more stable glycosidic
linkage and increased metabolic stability.4 Among these
nucleosides, -2Ј-deoxy-4Ј-thio-1Ј-pyrimidine nucleosides5
show potent anti-HSV (herpes simplex virus) and antitumor
activities, and -β-2Ј-deoxy-4Ј-thio-1Ј-purine nucleosides
exhibit potent anti-HBV (hepatitis
B virus) and anti-
HCMV (human cytomegalovirus) activity, but they were found
to be highly nephrotoxic when tested in vivo.6 Therefore, it was
interesting to synthesize -β-2Ј-deoxy-4Ј-thio-1Ј-purine nucleo-
sides and to compare their biological activities to those of their
corresponding -nucleosides, since many -nucleosides such as
3TC,7 -Fd4C,8 and -FMAU9 exhibit more potent antiviral
activities with much less cytotoxicity than their corresponding
 counterparts.
However, although -β-2Ј-deoxy-4Ј-thio-1Ј-purine nucleo-
sides show interesting biological activity, the biological activity
of β-2Ј-deoxy-4Ј-thio-1Ј-purine nucleosides has rarely been
reported in the literature because of the synthetic difficulties
in the 2-deoxy-4-thiosugar acetate, together with the unfavor-
able anomeric ratio produced during the condensation. For
example, when a -2-deoxy-4-thiosugar acetate was condensed
with purine bases, the α-anomer was always formed pre-
dominantly over the β-anomer (α:β = 9:1).10 Van Draanen
and co-workers6 obtained the desired β-anomer by a trans
glycosylation method using trans-N-deoxyribosylase, but this
method involved synthesis of an anomeric mixture of the -2Ј-
deoxy-4Ј-thio-1Ј-pyrimidine nucleoside followed by condens-
ation with purine bases in the presence of the transfer enzyme
(Scheme 1).
Therefore, it was necessary to develop a new efficient
synthetic method to obtain -β-2Ј-deoxy-4Ј-thio-1Ј-purine
nucleosides for the structure–activity relationship study, as
well as for the purpose of reducing the toxicity of -
nucleosides. Our laboratory developed a very short and efficient
route11 to the -4-thiosugar, in which the C2 position can
be modified selectively. Since neighboring group participation
by the C2 acyl group is a very efficient way to obtain the
β-anomer selectively in synthesizing nucleosides, our key
intermediate, an -4-thioarabitol derivative, was thought to be
an excellent synthon for the chemical synthesis of -2Ј-deoxy-
4Ј-thio-1Ј-purine nucleosides. Here, we report the highly
efficient synthesis of -β-2Ј-deoxy-4Ј-thio-1Ј-purine nucleo-
sides, utilizing the anchimeric effect of the C2 benzoyl group of
-4-thiosugar.
Scheme 1
Results and discussion
The glycosyl donor, -2-benzoyl-4-thiosugar acetate, was syn-
thesized from the -4-thioarabitol derivative 2,11 which could
be easily synthesized from 1,2-O-isopropylidene--xylose in
50% overall yield (Scheme 2). The benzoyl group of 2 was
DOI: 10.1039/b001240h
J. Chem. Soc., Perkin Trans. 1, 2000, 1327–1329
This journal is © The Royal Society of Chemistry 2000
1327

View Article Online
between 1Ј-H and 2Ј-H of 6ꢀ was smaller than that of 6ꢁ
(5.3%). It is of interest to note that the N-3 isomer [Rf = 0.32,
UV (MeOH) λmax 271 nm] was initially formed during the
condensation, and smoothly rearranged to the N-9 isomer
[Rf = 0.62, UV (MeOH) λmax 265 nm] on refluxing, as reported
by Chu and co-workers.14 Besides the UV data of 6ꢀ, the
assignment of N-9 regiochemistry was further confirmed by the
UV spectral data of the adenine analog 9 [UV (MeOH) λmax 260
nm] and N-methyladenine analog 10 [UV (MeOH) λmax 266
nm]. Treatment of 6ꢀ with sodium methoxide afforded 7, which
was deoxygenated using modified Barton’s conditions to give
the 2Ј-deoxy nucleoside 8. Treatment of 2Ј-deoxy derivative 8
with boron trichloride produced the 6-chloropurine derivative
9 in low yield because of the partial deprotection of the benzyl
group. However, use of an excess of boron tribromide gave
the desired 9 in 80% yield. Compound 9 was easily converted
to the adenine derivative 10 (95%) by treatment with meth-
anolic ammonia at 80 ЊC. The N-methyladenine analog 11 was
obtained by heating with 40% methylamine in methanol at
80 ЊC (96%). The antiviral assay of the final nucleosides 9–11 is
in progress and will be reported in due course.
Scheme 2
removed by treatment with sodium methoxide in quantitative
yield. To obtain the desired neighboring group effect during the
condensation reaction with nucleosidic base, the stereo-
chemistry of the C2 hydroxy group of 3 was inverted using
Mitsunobu conditions (PPh3, DEAD, benzoic acid, 60 ЊC) to
give 4 (50%) with the concomitant formation of compound 2
(10–20%), which resulted from the participation of the sulfur
atom of the 4-thiofuranose.12 The acetoxy group at the ano-
meric position was introduced by Pummerer rearrangement by
treating 4 with MCPBA, followed by refluxing with acetic
anhydride to yield an acetate 5 (65%).
In summary, we accomplished an efficient synthesis of -β-
2Ј-deoxy-4Ј-thio-1Ј-purine nucleosides through neighboring
group participation of the C2 benzoyl group of the -2-
benzoyl-4-thiosugar. This synthetic method illustrates the
general procedure for the predominant synthesis of β-2Ј-deoxy-
4Ј-thio-1Ј-purine nucleosides from our versatile intermediate,
1,4-anhydro-2-benzoyl-3,5-dibenzyl--4-thioarabitol.
The synthesis of the target nucleosides 9–11 is shown in
Scheme 3. Condensation of the acetate 5 with silylated 6-
Acknowledgements
This research was supported by a grant from the Korea
Research Foundation awarded in the Program Year 1997.
References and notes
1 M. Bobek and R. L. Whistler, J. Med. Chem., 1972, 15, 168.
2 M. R. Dyson, P. L. Coe and R. T. Walker, J. Chem. Soc., Chem.
Commun., 1991, 741.
3 (a) N. Ototani and R. L. Whistler, J. Med. Chem., 1974, 17, 535; (b)
J. A. Secrest III, K. N. Tiwari, J. M. Riordan and J. A. Montgomery,
J. Med. Chem., 1991, 34, 2361.
4 R. E. Parks, Jr., J. D. Stoeckler, C. Cambor, T. M. Savarese,
G. Crabtree and S.-H. Chu, Purine Nucleoside Phosphorylase and
5Ј-Methylthioadenosine Phosphorylase: Targets of Chemotherapy.
Molecular Actions and Targets for Cancer Chemotherapeutic Agents,
eds. A. C. Sartorelli, J. S. Lazo and J. R. Bertino, Academic Press,
New York, 1981, pp. 229–252.
5 (a) S. G. Rahim, N. Trivedi, M. V. Bogunovic-Batchelor, G. W.
Hardy, G. Mills, J. W. T. Selway, W. Snowden, E. Littler, P. L. Coe,
I. Basnak, R. F. Whale and R. T. Walker, J. Med. Chem., 1996,
39, 789; (b) M. R. Dyson, P. L. Coe and R. T. Walker, J. Med.
Chem., 1991, 34, 2782; (c) M. R. Dyson, P. L. Coe and R. T. Walker,
Carbohydr. Res., 1991, 216, 237; (d) J. Uenishi, K. Takahashi,
M. Motoyama, H. Akashi and T. Sasaki, Nucleosides Nucleotides,
1994, 13, 1347.
6 N. A. Van Draanen, G. A. Freeman, S. A. Short, R. Harvey,
R. Jansen, G. Szczech and G. W. Koszalka, J. Med. Chem., 1996, 39,
538.
7 S.-L. Doong, C.-H. Tasi, R. F. Schinazi, D. C. Liotta and Y.-C.
Cheng, Proc. Natl. Acad. Sci. USA, 1991, 88, 8495.
8 S.-H. Chen, S. Lin, I. King, T. Spinka, G. E. Dutschman, E. A.
Gullen, Y.-C. Cheng and T. W. Doyle, Bioorg. Med. Chem. Lett.,
1998, 8, 3245.
Scheme 3
9 T. Ma, S. B. Pai, Y. L. Zhu, J. S. Lin, K, Shanmuganathan, J. F. Du,
C. W. Wang, H. Kim, M. G. Newton, Y.-C. Cheng and C. K. Chu,
J. Med. Chem., 1996, 39, 2835.
10 J. A. Secrest III, K. N. Parker, K. N. Tiwari, L. Messini, S. C.
Shaddix, L. M. Rose, L. L. Bennett and J. A. Montgomery, Nucleo-
sides Nucleotides, 1995, 14, 675.
11 L. S. Jeong, H. R. Moon, Y. J. Choi, M. W. Chun and H. O. Kim,
J. Org. Chem., 1998, 57, 4821.
12 L. S. Jeong, S. J. Yoo, H. R. Moon, Y. H. Kim and M. W. Chun,
J. Chem. Soc., Perkin Trans. 1, 1998, 3325.
chloropurine in the presence of TMSOTf in 1,2-dichloroethane
gave the desired β-anomer 6ꢀ as the predominant product
(60%), which could be formed by the neighboring group effect
of the C2 benzoyl group, and also the α-anomer 6ꢁ as the minor
product (6%) after purification by silica gel column chrom-
atography.13 The assignments of the anomeric configurations of
6ꢁ and 6ꢀ were based on NOE experiments. When each 4Ј-H
peak of 6ꢁ and 6ꢀ was irradiated, enhancement of the 1Ј-H
peak of 6ꢀ was observed, suggesting the cis orientation, while
no enhancement of the 1Ј-H peak of 6ꢁ was observed, indicat-
ing the trans configuration. Additionally, the NOE effect (1.5%)
13 A solution of the acetate 5 (984 mg, 1.99 mmol) in 1,2-dichloroethane
(5 mL) was added to a solution of silylated 6-chloropurine (462 mg,
2.99 mmol) in 1,2-dichloroethane (2 mL) followed by the dropwise
1328
J. Chem. Soc., Perkin Trans. 1, 2000, 1327–1329

View Article Online
addition of trimethylsilyl trifluoromethanesulfonate (0.43 mL, 2.39
mmol) at 0 ЊC. The mixture was stirred at RT for 24 h. TLC gave a
slower moving spot due to the N-3 isomer (major product) which
was almost completely converted to the N-9 isomer (faster moving
spot) on refluxing for another 24 h. The usual work-up and
successive purification by silica gel chromatography (hexanes–
ethyl acetate = 4:1 to 3:1) gave 6β (Rf = 0.62, 703.8 mg, 60%)
C, 63.42; H, 4.64; N, 9.54. Found: C, 63.65; H, 4.54; N, 9.23%.
Compound 6α: 1H NMR δ 3.60 (dd, 1 H, J = 6.0 and 10.0 Hz,
5Ј-Ha), 3.65 (dd, 1 H, J = 6.3 and 10.0 Hz, 5Ј-Hb), 4.06–4.11 (dt,
1 H, J = 3.2 and 6.0 Hz, 4Ј-H), 4.43 (t, 1 H, J = 3.8 Hz, 3Ј-H), 4.58
(s, 2 H, CH2C6H5), 4.59 (s, 2 H, CH2C6H5), 5.90 (dd, 1 H, J = 3.8 and
6.1 Hz, 2Ј-H), 6.57 (d, 1 H, J = 6.1 Hz, 1Ј-H), 7.18–7.58 (m, 15 H,
3 × C6H5), 8.51 (s, 1 H, H-8), 8.83 (s, 1 H, H-2). Calcd for
C31H27ClN4O4S: C, 63.42; H, 4.64; N, 9.54. Found: C, 63.33; H, 4.76;
N, 9.14%.
1
and 6α (Rf = 0.55, 70.1 mg, 6%). Compound 6β: H NMR δ 3.82–
3.89 (m, 2 H, 5Ј-H), 3.91–3.93 (m, 1 H, 4Ј-H), 4.50–4.68 (m, 5 H,
3Ј-H and 2 × CH2C6H5), 6.04 (dd, 1 H, J = 3.7 and 5.2 Hz, 2Ј-H),
6.45 (d, 1 H, J = 5.2 Hz, 1Ј-H), 7.22–8.07 (m, 15 H, 3 × C6H5),
8.76 (s, 1 H, H-8), 8.77 (s, 1 H, H-2). Calcd for C31H27ClN4O4S:
14 H. O. Kim, R. F. Schinazi, S. Nampalli, K. Shanmuganathan, D. L.
Cannon, A. J. Alves, L. S. Jeong, J. W. Beach and C. K. Chu, J. Med.
Chem., 1993, 36, 30.
J. Chem. Soc., Perkin Trans. 1, 2000, 1327–1329
1329