(2R,3S,5S)-2-Acetoxy-3-fluoro-5-(p-toluoyloxymethyl)tetrahydrofuran: A key intermediate for the practical synthesis of 9-(2,3-dideoxy-2-fluoro-β-D-threo-pentofuranosyl)adenine (FddA)

Article abstract of DOI:10.1016/S0040-4039(01)00855-3

A highly efficient synthesis of (2R,3S,5S)-2-acetoxy-3-fluoro-5-(p-toluoyloxymethyl)tetrahydrofuran (2a) was developed. As a key intermediate, 2a was effectively applied to the synthesis of 9-(2,3-dideoxy-2-fluoro-β-D-threo-pentofuranosyl)adenine (FddA), thus, providing a practical synthetic approach to FddA.

Full text of DOI:10.1016/S0040-4039(01)00855-3

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 42 (2001) 4787–4789
(2R,3S,5S)-2-Acetoxy-3-fluoro-5-(p-toluoyloxymethyl)tetrahydro-
furan: a key intermediate for the practical synthesis of
9-(2,3-dideoxy-2-fluoro-b- -threo-pentofuranosyl)adenine (FddA)
D
Fuqiang Jin,* Dengjin Wang, Pat N. Confalone, Michael E. Pierce, Zhe Wang, Guoyou Xu,
Anusuya Choudhury and Dieu Nguyen
Chemical Process R&D, The DuPont Pharmaceuticals Company, DuPont Experimental Station, Wilmington,
DE 19880-0336, USA
Received 18 April 2001; revised 11 May 2001; accepted 16 May 2001
Abstract—A highly efficient synthesis of (2R,3S,5S)-2-acetoxy-3-fluoro-5-(p-toluoyloxymethyl)tetrahydrofuran (2a) was devel-
oped. As a key intermediate, 2a was effectively applied to the synthesis of 9-(2,3-dideoxy-2-fluoro-b-D-threo-pentofuranos-
yl)adenine (FddA), thus, providing a practical synthetic approach to FddA. © 2001 Dupont Pharmaceuticals Company.
Published by Elsevier Science Ltd. All rights reserved.
NH2
9-(2,3-Dideoxy-2-fluoro-b-D-threo-pentofuranosyl)ade-
PO
O
F
nine (FddA, 1) is a novel drug candidate for the
treatment of AIDS.1–3 One major developmental issue
of this compound is the lack of a practical, commercial
process. Although extensive efforts have been devoted
to the discovery of new synthetic approaches,1,2,4–9 the
challenge of identifying a scalable process for FddA
still remains. Several methods for the synthesis of FddA
have been reported based on the modification of non-
fluorinated nucleosides such as cordycepin.4,5 However,
the ineffective introduction of the fluorine atom in the
required b-orientation of the aglycone renders these
methods impractical for a large scale synthesis. Though
attractive in principal, more convergent approaches2,6,10
consisting of coupling a fluorosugar, especially a 3-
deoxy fluorosugar (2), with the purine base remain
problematic due to the inefficient coupling reaction and
the difficult synthesis of the fluorosugar. Our intensive
research efforts on this problem have led to the devel-
opment of a scalable process for FddA based on the
key intermediate 2a. Herein, we would like to report
our results.
N
N
N
HO
X
N
O
F
2
2a: P = p-CH3C6H4CO
1
X = OAc
Patrick et al. reported a synthesis of (S)-(−)-2-fluoro-4-
(hydroxymethyl)-2-buten-4-olide (3)11 and its applica-
tion to the preparation of 1-(2%-fluoro-2%,3%-dideoxy-
b-D
-threo-pentofuranosyl)thymine.12 Recognizing the
potential of 3 as a feasible precursor of a variety of
3-deoxy fluorosugars 2 and the potential key role a
suitable 3-deoxy fluorosugar might play in the FddA
process development, we were attracted to the conver-
sion of 3 into FddA through the intermediacy of a
reliable 3-deoxy fluorosugar. To this end, we have
developed a highly efficient synthesis of 3-deoxy fluoro-
sugar 2a (Scheme 1).
Toluoylation of the fluoroolefin 3 gave 4 as a crystalline
solid in an excellent yield (95%). The highly crystalline
nature of 4 made it possible to prepare this compound
directly from commercially available 1,2:5,6-di-O-iso-
propylidene-D-mannitol without purifying any interme-
diates in a yield of 65%. Compound 4 was hydro-
genated to 513 in the presence of Pd-C in quantitative
yield but, surprisingly, with a diastereoselectivity which
Keywords: coupling reactions; nucleosides; lactones; purines.
* Corresponding author. Tel.: (302) 695-8666; fax: (302) 695-4632;
e-mail: fuqiang.jin@dupontpharma.com
0040-4039/01/$ - see front matter © 2001 Dupont Pharmaceuticals Company. Published by Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(01)00855-3

4788
F. Jin et al. / Tetrahedron Letters 42 (2001) 4787–4789
HO
p-TolO
p-TolO
O
O
O
F
a
b
O
O
O
95%
97%
F
F
3
4
5
p-TolO
p-TolO
O
F
O
F
d
c
OH
OAc
93%
100%
2a
6
p-Tol = 4-CH3C6H4CO
Scheme 1. Reagents and conditions: (a) p-TolCl, Py, CH2Cl2, 0°C–rt; (b) 5% Pd-C, H2 (10 psi), EtOAc, rt; (c) red-Al,
2-hydroxypyridne, THF, −30°C; (d) Ac2O, cat. conc. H2SO4, CH2Cl2, rt.
is dependent on the hydrogen pressure. When the
hydrogenation was conducted at a hydrogen pressure of
40 psi, a moderate diastereoselectivity (79.5% de) was
observed; however, at 20 psi the diastereoselectivity was
increased to 90%. A satisfactorily high diastereosel-
ectivity (98%) was achieved at a H2 pressure of 10 psi.
The structure of 5 was determined by single-crystal
X-ray crystallography.
conditions using MeOH or H2O. To selectively reduce 5
to the desired lactol 6, a wide array of reducing
reagents with varying degrees of reducing ability such
as NaBH4,15 L-Selectride,16 red-Al,17 disiamylborane,18
Vitride solution19 and LiAlH(Ot-Bu)3 were screened
20
to effect this reduction, but all failed to give satisfactory
results. We reasoned that a selective reducing reagent
should possess the ability to efficiently stabilize the
lactoxide intermediate so that an over-reduction can be
prevented. Thinking along this line, we tried a combin-
ation of LAH with various alcohols of varying pKas
such as phenol, MeOH, EtOH, CF3CH2OH, aiming to
lower the reducing ability of LAH while, in the mean-
time, to increase its coordination ability with the lactox-
ide. As expected, we observed that the lower the pKa of
the alcohol, the more selective its combination with
LAH. When the alcohol was phenol and at a reaction
temperature of −78°C, 6 was obtained in an isolated
yield of 93% with less than 4% of by-product 7. Excel-
lent selectivity can also be achieved by the combination
of LAH with other alcohols such as hydroxypyridine,
or combination of other aluminum hydride such as
red-Al with alcohols. A more impressive combination
which offers practical advantage over the combination
of LAH and phenol is that of red-Al with 2-hydroxy-
pyridine. This reducing reagent allows 5 to be con-
verted into 6 at higher temperature (−30°C) with a
comparable efficiency and selectivity. Finally, a quanti-
Although reduction of a g-lactone to the corresponding
lactol is a common transformation,14–20 conversion of 5
into 6 was initially found to be problematic. When 5
was treated with DIBAL-H at −78°C in CH2Cl2, the
desired lactol 6 was obtained in a yield of only 40%
after quenching with MeOH. A significant amount of
the over-reduction product 7 (30%) was isolated along
with the tetrahydropyran 8 (15%), which was character-
ized by conversion into its corresponding acetate 9.
This pyranose presumably results from migration of the
toluoyl group. Indeed, investigation into the formation
of pyranose 8 indicated that it was mainly generated
during the work-up. It was further observed that expo-
sure of 6 to bases such as Et3N or aqueous NaOH leads
to a partial conversion into 8, suggesting that 8 results
from a base-induced toluoyl group migration of 6.
Based on these observations, we eliminated the forma-
tion of 8 by quenching the reduction under acidic
conditions with acetic acid rather than under basic
NHBz
p-TolO
N
N
O
F
N
N
p-TolO
p-TolO
N
N
N
f
O
F
e
O
F
2a
+
N
68%
100%
Br
NHBz
10a
10b
HO
O
F
N
g
N
+
N
FddA
66%
N
NH2
11
Scheme 2. Reagents and conditions: (e) AcOH–HBr, CH2Cl2; (f) N-benzoylaminopurine, NaH, THF, reflux; (g) MeONa, MeOH,
reflux.

F. Jin et al. / Tetrahedron Letters 42 (2001) 4787–4789
4789
tative conversion of 6 into 2a was achieved by treat-
2. Marquez, V. E.; Tseng, C. K.-H.; Mitsuya, H.; Aoki, S.;
Kelly, J. A.; Ford, Jr., H.; Roth, J. S.; Broder, S.; Johns,
D. G.; Driscoll, J. S. J. Med. Chem. 1990, 33, 978.
3. Graul, A.; Silvestre, J.; Castaner, J. Drug of the Future
1998, 23, 1176.
ment with Ac2O and catalytic concentrated H2SO4.
O
OH
F
O
OAc
F
OH
F
p-TolO
OH
p-TolO
p-TolO
4. Herdewijn, P.; Pauwels, R.; Baba, M.; Balzarini, J.; De
Clercq, E. J. Med. Chem. 1987, 30, 2131.
7
8
9
5. Takamatsu, S.; Maruyama, T.; Katayama, S.; Hirose, N.;
Naito, M.; Izawa, K. Tetrahedron Lett. 2001, 42, 2325.
6. Wysocki, Jr., R. J.; Siddiqui, M. A.; Barchi, Jr., J. J.;
Driscoll, J. S.; Marquez, V. E. Synthesis 1991, 1005.
7. Shiragami, H.; Tanaka, Y.; Uchida, Y.; Iwagami, H.;
Izawa, K.; Yukawa, T. Nucleosides Nucleotides 1992, 11,
391.
8. Marquez, V. E.; Driscoll, J. S.; Wysocki, Jr., R. J.;
Siddiqui, M. A. (Dept. Health Human Serv. [USA]). US
5817799.
9. Siddiqui, M. A.; Driscoll, J. S.; Marquez, V. E. Tetra-
hedron Lett. 1998, 39, 1657.
10. Siddiqui, M. A.; Marquez, V. E.; Driscoll, J. S.; Barchi,
Jr., J. J. Tetrahedron Lett. 1994, 35, 3263.
11. Patrick, T. B.; Lanahan, M. V.; Yang, C.; Walker, J. K.;
Hutchinson, C. L.; Neal, B. E. J. Org. Chem. 1994, 59,
1210.
With a highly efficient and economically feasible syn-
thesis of 2a developed, we then studied its conversion to
FddA as shown in (Scheme 2). Coupling of a 3-
deoxyfluorosugar with a purine base has been plagued
by low yields and poor b/a selectivity, necessitating a
tedious separation of the desired b anomer.6 To
improve the b/a selectivity, a 3-deoxyfluorosugar is
usually converted into its corresponding 1-a-halosugar
which is then coupled with a base under conditions
favoring an SN2 nucleophilic displacement. To define a
process for coupling 2a with a purine base suitable for
large-scale synthesis of FddA, we screened different
derivatives of purine bases. For the coupling reaction,
2a was first converted into the corresponding a-bromo-
sugar stereoselectively in quantitative yield by treat-
ment with HBr/HOAc. The bromosugar was then
reacted with a purine base either as its persilylated
analog or the sodium salt. We found that coupling of
the sodium salt of 6-N-benzoylpurine with 2a in THF
gave a cleaner reaction and better b/a selectivity than
using purine or 6-chloropurine. Thus, after conversion
into its corresponding a-bromide, 2a was treated with
the sodium salt of 6-N-benzoylpurine generated by
reaction of 6-N-benzoylpurine with NaH in refluxing
THF to afford the desired coupling product 10a and its
anomer 10b in a yield of 68% and with a b/a ratio of 8
to 1. Although separation of 10a from 10b at this stage
proved difficult, a facile recrystallization after hydroly-
sis easily yielded pure FddA free of the corresponding
hydrolysis product of 10b, the a-anomer of FddA (11).
In this manner, we converted 2a into FddA in an
overall yield of 45% after recrystallization.
12. Patrick, T. B.; Ye, W. J. Fluor. Chem. 1998, 90, 53.
13. Spectral data for compound 5 and its a-epimer. (3R,5S)-
3-Fluoro-5-(p-toluoyloxymethyl)-2-tetrahydrofuranone
1
(5): white solid, mp 138–140°C; H NMR (CDCl3) l 2.39
(m, 1H), 2.40 (s, 3H), 2.86 (m, 1H), 4.42 (dd, J=11.0,
4.8, 1H), 4.60 (dd, J=12.0, 3.0, 1H), 4.80 (m, 1H), 5.30
(dt, J=50.9, 8.8, 1H), 7.24 (d, J=8.1, 2H), 7.92 (d,
J=8.1, 2H); 19F NMR (CDCl3) l −193.2 (m); MS (ESI)
m/z (relative intensity %), 253 (M+1, 100); Anal. calcd
for C13H13FO4: C, 61.90; H, 5.20. Found: C, 61.90; H,
5.13.
(3S,5S)-3-Fluoro-5-(p-toluoyloxymethyl)-2-tetra-
hydrofuranone (a-epimer of 5): white solid, mp 68–70°C.
1H NMR (CDCl3) l 2.44 (s, 3H), 2.57–2.76 (m, 2H), 4.52
(m, 2H), 5.06 (m, 1H), 5.36 (ddd, J=52.0, 7.3, 6.6, 1H),
7.27 (d, J=8.1, 2H), 7.86 (d, J=8.1, 2H); 19F NMR
(CDCl3) l −190.3 (m); MS (ESI) m/z (relative intensity
%), 253 (M+1, 100). Anal. calcd for C13H13FO4: C, 61.90;
H, 5.20. Found: C, 61.85; H, 5.19.
In conclusion, we have developed a highly efficient
synthesis of the 3-deoxy fluorosugar 2a and converted it
14. Kim, C. U.; Misco, P. F. Tetrahedron Lett. 1992, 33,
5733.
15. Wolfrom, M. L.; Anno, K. J. Am. Chem. Soc. 1952, 74,
5883.
16. Kotra, L. P.; Newton, M. P.; Chu, C. K. Carbohydr. Res.
1998, 306, 69.
17. Lieb, F.; Niewohner, U.; Wendish, D. Liebigs Ann.
Chem. 1987, 609.
.
into FddA via an efficient process. The high efficiency
in each step and simplicity of the purification of the
intermediates and the final product render this route
practical for large-scale synthesis of FddA. In addition,
the intermediate 2a should prove to be a useful building
block for the synthesis of other 3-deoxyfluoro
nucleosides.
18. Pedersen, C.; Jensen, H. S. Acta Chem. Scand. 1994, 48,
222.
19. Kanazawa, K.; Kotsuki, H.; Tokoroyama, T. Tetra-
hedron Lett. 1975, 3651.
20. Kotra, L. P.; Newton, M. G.; Chu, C. K. Carbohydr.
Res. 1998, 306, 69.
References
1. Marquez, V. E.; Tseng, C. K.-H.; Kelly, J. A.; Mitsuya,
H.; Broder, S.; Roth, J. S.; Driscoll, J. S. Biochem.
Pharmacol. 1987, 36, 2719.