2'-chloro-2',3'-dideoxy-3'-fluoro-d-ribonucleosides: synthesis, stereospecificity, some chemical transformations, and conformational analysis.

Article abstract of DOI:10.1021/jo0340859

The synthesis of methyl 5-O-benzoyl-2-chloro-2,3-dideoxy-3-fluoro-beta-d-ribofuranoside (5) and its use as a glycosylating agent for persilylated thymine, N(6)-benzoyladenine, and N(4)-benzoylcytosine are described (Scheme 1). The 2'-chloro-2',3'-dideoxy-3'-fluoro-d-ribonucleosides 10-12 synthesized were transformed to 2',3'-dideoxy-3'-fluoro-alpha- and -beta-d-erythro-pentofuranoside nucleosides of thymine (13a,b), adenine (14a,b), and cytidine (15a,b) by treatment with tributyltin hydride in the presence of alpha,alpha'-azobisisobutyronitrile (Scheme 2). Treatment of 2'-chloro-2',3'-dideoxy-3'-fluoro-d-ribonucleosides with 1 M MeONa/MeOH under reflux for 1-5 h afforded 2',3'-didehydro-2',3'-dideoxy-2'-chloro-d-pentofuranosyl nucleosides as the principal products (47-81%) of the reaction, along with recovered starting nucleoside (11-33%) (Scheme 3). Easy HF elimination was also observed in the case of the 2'-azido-2',3'-dideoxy-3'-fluoro-beta-d-ribofuranosides of thymine (17) and adenine (20) (Scheme 3). The role of conformational peculiarities of 2'-chloro-2',3'-dideoxy-3'-fluoro-d-ribonucleosides as well as of 17 and 20 in the observed exclusive elimination of HF is discussed. The conformational analysis of a rather broad palette of 2,3-dideoxy-3-fluoro-2-(X-substituted)-d-ribofuranosides was performed with the aid of the PSEUROT (version 6.3) program, using (i) the recently reparametrized Karplus-type relation (Chattopadhyaya and co-workers. J. Org. Chem. 1998, 63, 4967) and (ii) empirical bond angle correction terms suggested by us. The predictive power of the Brunck and Weinhold model (J. Am. Chem. Soc. 1979, 101, 1700) of the gauche effect between atoms and groups as a conformational driving force acting upon the pentofuranose ring is explored. Their model invokes maximum antiperiplanar sigma <--> sigma stabilization when the donating bond is the least polar one and the acceptor orbital is at the most polarized bond and is found at least as satisfactory, and in various specific cases more so than, as rationalizations on the basis of the preference of the gauche vs the trans conformation of two vicinal electronegative substituents (Wolfe. Acc. Chem. Res. 1972, 5, 102).

Full text of DOI:10.1021/jo0340859

2′-Ch lor o-2′,3′-d id eoxy-3′-flu or o-D-r ibon u cleosid es: Syn th esis,
Ster eosp ecificity, Som e Ch em ica l Tr a n sfor m a tion s, a n d
Con for m a tion a l An a lysis
Igor A. Mikhailopulo,†,‡ Tamara I. Pricota,† Grigorii G. Sivets,† and Cornelis Altona*,§
Institute of Bioorganic Chemistry, National Academy of Sciences, 220141 Minsk, Belarus, and Gorlaeus
Laboratories, Leiden Institute of Chemistry, Leiden University, NL-2300 RA Leiden, The Netherlands
c.altona@chem.leidenuniv.nl
Received J anuary 24, 2003
The synthesis of methyl 5-O-benzoyl-2-chloro-2,3-dideoxy-3-fluoro-â-D-ribofuranoside (5) and its
use as a glycosylating agent for persilylated thymine, N6-benzoyladenine, and N4-benzoylcytosine
are described (Scheme 1). The 2′-chloro-2′,3′-dideoxy-3′-fluoro-D-ribonucleosides 10-12 synthesized
were transformed to 2′,3′-dideoxy-3′-fluoro-R- and -â-D-erythro-pentofuranoside nucleosides of
thymine (13a ,b), adenine (14a ,b), and cytidine (15a ,b) by treatment with tributyltin hydride in
the presence of R,R′-azobisisobutyronitrile (Scheme 2). Treatment of 2′-chloro-2′,3′-dideoxy-3′-fluoro-
D-ribonucleosides with 1 M MeONa/MeOH under reflux for 1-5 h afforded 2′,3′-didehydro-2′,3′-
dideoxy-2′-chloro-D-pentofuranosyl nucleosides as the principal products (47-81%) of the reaction,
along with recovered starting nucleoside (11-33%) (Scheme 3). Easy HF elimination was also
observed in the case of the 2′-azido-2′,3′-dideoxy-3′-fluoro-â-D-ribofuranosides of thymine (17) and
adenine (20) (Scheme 3). The role of conformational peculiarities of 2′-chloro-2′,3′-dideoxy-3′-fluoro-
D-ribonucleosides as well as of 17 and 20 in the observed exclusive elimination of HF is discussed.
The conformational analysis of a rather broad palette of 2,3-dideoxy-3-fluoro-2-(X-substituted)-D-
ribofuranosides was performed with the aid of the PSEUROT (version 6.3) program, using (i) the
recently reparametrized Karplus-type relation (Chattopadhyaya and co-workers. J . Org. Chem.
1998, 63, 4967) and (ii) empirical bond angle correction terms suggested by us. The predictive power
of the Brunck and Weinhold model (J . Am. Chem. Soc. 1979, 101, 1700) of the gauche effect between
atoms and groups as a conformational driving force acting upon the pentofuranose ring is explored.
Their model invokes maximum antiperiplanar σ T σ* stabilization when the donating bond is the
least polar one and the acceptor orbital is at the most polarized bond and is found at least as
satisfactory, and in various specific cases more so than, as rationalizations on the basis of the
preference of the gauche vs the trans conformation of two vicinal electronegative substituents (Wolfe.
Acc. Chem. Res. 1972, 5, 102).
In tr od u ction
HIV activity has been reported for 1-(2,3-dideoxy-3-fluoro-
â-D-erythro-pentofuranosyl)thymine (FLT, 13b)5-7 and
1-(2,3-dideoxy-3-fluoro-â-D-erythro-pentofuranosyl)-5-chlo-
rouracil;6,7 the former was found to be the most potent
anti-HIV-1 nucleoside analogue, whereas the latter
emerged as the most selective inhibitor of HIV-1 replica-
tion.
A number of 2′,3′-dideoxy-3′-fluoro-â-D-erythro-pento-
furanosyl nucleosides display high antiviral activity (for
reviews, see, e.g., refs 1-3; see also refs 4-9). Potent anti-
* To whom correspondence should be addressed.
† National Academy of Sciences.
‡ Present address: University of Kuopio, Department of Pharma-
ceutical Chemistry, P.O. Box 1627, Kuopio, FIN-70211, Finland.
§ Leiden University.
Since the first synthesis of FLT from thymidine,10
much attention has been devoted to the preparation of
2′,3′-dideoxy-3′-fluoro-â-D-erythro-pentofuranosyl nucleo-
sides (earlier works are reviewed in ref 11). Two ap-
proaches are used for the preparation of these analogues,
(1) Langen, P. Antimetabolites of Nucleic Acid Metabolism; Gordon
& Breach: New York, 1975; pp 177-178.
(2) De Clercq, E. AIDS Res. Hum. Retroviruses 1992, 8, 119-134.
(3) Marquez, V. A.; Lim, B. B.; Barchi, J . J .; Nicklaus, M. C.
Conformational studies and anti-HIV activity of mono- and difluo-
rodideoxy nucleosides. In Nucleosides and Nucleotides as Antitumor
and Antiviral Agents; Chu, C. K., Baker, D. C., Eds.; Plenum Press:
New York and London, 1993; pp 265-284.
(7) Herdewijn, P. A. M.; Van Aerschot, A.; Balzarini, J .; Pauwels,
R.; De Clercq, E.; Everaert, D. H.; De Winter, H. L.; Peeters, O. M.;
Blaton, N. M.; De Ranter, C. J . Med. Chem. Res. 1991, 1, 9-19.
(8) Hafkemeyer, P.; Keppler-Hafkemeyer, A.; al Haya, M. A.; von
J anta-Lipinski, M.; Matthes, E.; Lehmann, C.; Offensperger, W.-B.;
Offensperger, S.; Gerok, W.; Blum, H. E. Antimicrob. Agents Chemo-
ther. 1996, 40, 792-794.
(9) Mentel, R.; Kinder, M.; Wegner, U.; von J anta-Lipinski, M.;
Matthes, E. Antiviral Res. 1997, 34, 113-119.
(10) Etzold, G.; Hintsche, R.; Kowollik, G.; Langen, P. Tetrahedron
1971, 27, 2463-2472.
(4) Pauwels, R.; Baba, M.; Balzarini, J .; Herdewijn, P.; Desmyter,
J .; Robins, M. J .; Zou, R.; Madej, D.; De Clercq, E. Biochem. Pharmacol.
1988, 37, 1317-1325.
(5) Hartmann, H.; Vogt, M. W.; Durno, A. G.; Hirsch, M. S.;
Hunsmann, G.; Eckstein, F. AIDS Res. Hum. Retroviruses 1988, 4,
457-466.
(6) Van Aerschot, A.; Herdewijn, P.; Balzarini, J .; Pauwels, R.; De
Clercq, E. J . Med. Chem. 1989, 32, 1743-1749.
10.1021/jo0340859 CCC: $25.00 © 2003 American Chemical Society
Published on Web 07/02/2003
J . Org. Chem. 2003, 68, 5897-5908
5897

Mikhailopulo et al.
viz., (i) transformation of natural 2′-deoxyribonucleosides
or less serviceable ribonucleosides to the desired dideoxy-
fluoro derivatives11-15 and (ii) coupling of heterocyclic
bases with suitable furanose derivatives containing the
C3-fluorine atom or the C3-xylo-hydroxyl group, which
may be replaced by a fluorine atom6,11,16-21 (for a recent
review, see ref 22). The most important advantage of the
first approach lies in the presence of the requisite
â-configuration at the anomeric center. Nonetheless, the
preparation of FLT from thymidine through the inter-
mediary O2,3′-anhydro derivative is probably the only
expedient example.23
The main limitation of the second approach is the poor
stereoselectivity of the glycosidic bond formation employ-
ing suitable derivatives of 2,3-dideoxy-3-fluoro-D-erythro-
pentofuranose as glycosylating agents. On the other
hand, the use of the universal glycosylating agents makes
possible the preparation of a rather broad spectrum of
nucleosides having both natural and modified heterocy-
clic bases. The transglycosylation reaction employing the
5′-O-acylated FLT as glycosyl donor and purine bases as
glycosyl acceptors24 combines both of the above ap-
proaches.
essential improvement of the â-stereoselectivity of the
convergent methods. The elegant stereocontrolled syn-
theses of pyrimidine 2′,3′-dideoxy-â-D-nucleosides includ-
ing AZT and FLT based on the intramolecular glycosyl-
ation concept have recently been described.32-34
As a continuation of our efforts in the synthesis of 2,3-
dideoxy-3-fluoro-D-erythro-pentofuranosides,12,18,20,24 we
describe a new approach for the preparation of these
compounds by coupling methyl 5-O-benzoyl-2-chloro-2,3-
dideoxy-3-fluoro-â-D-ribofuranoside (5) with persilylated
thymine, N6-benzoyladenine, and N4-benzoylcytosine fol-
lowed by the radical-mediated hydrodehalogenation of
the prepared nucleosides and subsequent deprotection.
We also demonstrate that the dehalogenation of 2′-chloro-
2′,3′-dideoxy-3′-fluoro-D-ribonucleosides with 1 M MeONa/
MeOH under reflux for 1-5 h afforded 2′,3′-didehydro-
2′,3′-dideoxy-2′-chloro-D-pentofuranosyl nucleosides as
the principal products. Similar easy elimination of HF
was also observed in the case of 2′-azido-2′,3′-dideoxy-
3′-fluoro-â-D-ribofuranosides of thymine (17) and adenine
(20). The role of conformational peculiarities of 2′-chloro-
2′,3′-dideoxy-3′-fluoro-D-ribonucleosides as well as 17 and
20 in the observed exclusive elimination of HF is dis-
cussed.
The use of pentofuranosyl sugars containing at the C2-
atom an R-arranged iodine,25,26 phenylsulfenyl,27,28 phen-
ylselenenyl,21,29 or m-trifluoromethylbenzoyl group30,31 as
the transient anomeric control group, which can be
removed after the glycosidic bond formation, led to an
The conformational behavior of the various pentofura-
noses is discussed in terms of the anomeric effect (AE)
and an alternative model of the gauche effect (GE).
Resu lts a n d Discu ssion
(11) Herdewijn, P.; Van Aerschot, A.; Kerremans, L. Nucleosides
Nucleotides 1989, 8, 65-96.
Ch em ica l Tr a n sfor m a tion s. The key glycosylating
riboside 5 was prepared from methyl 3-deoxy-5-O-benzyl-
3-fluoro-2-O-tosyl-â-D-arabinofuranoside (1)35 in three
steps using two different sequences of the same chemical
reactions (Scheme 1). Two-step substitution of the ben-
zoyl group for the benzyl group followed by replacement
of the tosyloxy group by a chlorine atom gave the riboside
5 in 51% combined yield. The reversed sequence of the
chemical transformations resulted in a somewhat lower
yield (45%) of the desired 5. The reaction of the riboside
5 with persilylated thymine in the presence of trimeth-
ylsilyl triflate [(TMS)Tfl] (1.0:2.0:3.0, mol) in refluxing
acetonitrile for 5 h followed by chromatography afforded
a mixture of the R- and â-anomers 7a ,b (the ratio of R to
â was 1:2 according to 1H NMR) in 49% yield and a
mixture of the methyl glycosides 5 and 6 in 19% yield.
This result revealed a close resemblance with the previ-
ously described coupling of methyl 2-azido-5-O-benzoyl-
2,3-dideoxy-3-fluoro-â-D-ribofuranoside with persilylated
thymine under similar conditions.35 Treatment of 7a ,b
with methanolic ammonia and subsequent chromatog-
raphy on silica gel afforded 1-(2-chloro-2,3-dideoxy-3-
fluoro-â-D-ribofuranosyl)thymine (10b) and its R-anomer
10a in 55% and 28% yield, respectively.
(12) Zaitseva, G. V.; Kvasyuk, E. I.; Poopeiko, N. E.; Kulak, T. I.;
Pashinnik, V. E.; Tovstenko, V. I.; Markovskii, L. M.; Mikhailopulo, I.
A. Bioorg. Chem. (Moscow) 1988, 14, 1275-1281; Chem. Abstr. 1989,
111, 7722.
(13) Green, K.; Blum, D. M. Tetrahedron Lett. 1991, 32, 2091-2094.
(14) Seela, F.; Muth, H.-P. Helv. Chim. Acta 1991, 74, 1081-1090.
(15) Marchand, A.; Mathe, C.; Imbach, J .-L.; Gosselin, G. Nucleo-
sides, Nucleotides, Nucleic Acids 2000, 19, 205-217.
(16) Fleet, G. W. J .; Son, J . C.; Derome, A. E. Tetrahedron 1988,
44, 625-636.
(17) Motawia, M. S.; Pedersen, E. B. Liebigs Ann. Chem. 1990,
1137-1139.
(18) Mikhailopulo, I. A.; Pricota, T. I.; Poopeiko, N. E.; Klenitskaya,
T. V.; Khripach, N. B. Synthesis 1993, 700-704.
(19) Hager, M. W.; Liotta, D. C. Tetrahedron Lett. 1992, 33, 7083-
7086.
(20) Poopeiko, N. E.; Khripach, N. B.; Kazimierczyk, Z.; Balzarini,
J .; De Clercq, E.; Mikhailopulo, I. A. Nucleosides Nucleotides 1997,
16, 1083-1086.
(21) Poopeiko, N.; Fernandez, R.; Barrena, M. I.; Castillon, S.;
Fornies-Camer, J .; Cardin, C. J . J . Org. Chem. 1999, 64, 1375-1379.
(22) Viani, F. In Enantiocontrolled Synthesis of Fluoro-Organic
Compounds: Stereochemical Challenges and Biomedical Targets;
Soloshonok, V. A., Ed.; J ohn Wiley & Sons Ltd.: New York, 1999;
Chapter 13, pp 419-449.
(23) Huang, J .-T.; Chen, L.-C.; Wang, L.; Kim, M.-H.; Warshaw, J .
A.; Armstrong, D.; Zhu, Q.-Y.; Chou, T.-C.; Watanabe, K. A.; Matulic-
Adamic, J .; Su, T.-L.; Fox, J . J .; Polsky, B.; Baron, P. A.; Gold, J . W.
M.; Hardy, W. D.; Zuckerman, E. J . Med. Chem. 1991, 34, 1640-1646.
(24) Kvasyuk, E. I.; Zaitseva, G. V.; Savochkina, L. P.; Chidgeavadze,
Z. G.; Beabealashvilli, R. Sh.; Langen, P.; Mikhailopulo, I. A. Bioorg.
Chem. (Moscow) 1989, 15, 781-795; Chem. Abstr. 1990, 112, 3274.
(25) Kim, C. U.; Misco, P. F. Tetrahedron Lett. 1992, 33, 5733-5736.
(26) McDonald, F. E.; Gleason, M. M. Angew. Chem., Int. Ed. Engl.
1995, 34, 350-352.
Along the same line, the condensation of silylated N6-
benzoyladenine with methyl glycoside 5 in the presence
of excess SnCl4 (2.0:1.0:5.0, mol) in a refluxing acetoni-
(27) Wang, J .; Wurster, J . A.; Wilson, L. J .; Liotta, D. C. Tetrahedron
Lett. 1993, 34, 4881-4884.
(28) Kawakami, K.; Ebata, T.; Kozeki, K.; Matsushita, H.; Naoi, Y.;
Itoh, K. Chem. Lett. 1990, 8, 1459-1462.
(32) J ung, M. E.; Castro, C. J . Org. Chem. 1993, 58, 807-808.
(33) Sujino, K.; Sugimura, H. Tetrahedron Lett. 1994, 35, 1883-
1886.
(34) Lipshnitz, B. H.; Stevens, K. L.; Lowe, R. F. Tetrahedron Lett.
1995, 36, 2711-2712.
(35) Mikhailopulo, I. A.; Poopeiko, N. E.; Pricota, T. I.; Sivets, G.
G.; Kvasyuk, E. I.; Balzarini, J .; De Clercq, E. J . Med. Chem. 1991,
34, 2195-2202.
(29) Beach, J . W.; Kim, H. O.; J eong, L. S.; Nampalli, S.; Islam, Q.;
Ahn, S. K.; Babu, J . R.; Chu, C. K. J . Org. Chem. 1992, 57, 3887-
3894.
(30) Huang, Z.; Schneider, K. C.; Benner, S. A. J . Org. Chem. 1991,
56, 3869-3882.
(31) Park, M.; Rizzo, C. J . J . Org. Chem. 1996, 61, 6092-6093.
5898 J . Org. Chem., Vol. 68, No. 15, 2003

2′-Chloro-2′,3′-dideoxy-3′-fluoro-D-ribonucleosides
SCHEME 1a
SCHEME 2
a
(A) Nucleoside(s)/Bu3SnH/AIBN (1.0:6.0:0.33, mol), dioxane.
(B) Nucleoside/Bu3SnH/AIBN (1.0:5.0:0.33, mol), dioxane/MeOH
(1:1, vol). (C) NH3/MeOH, rt, 24-48 h.
(1:7) vs the similar reaction of methyl 2-azido-5-O-
benzoyl-2,3-dideoxy-3-fluoro-â-D-ribofuranoside (ca. 50%
combined, R:â ratio 1:4). Treatment of protected indi-
vidual nucleosides 8a ,b and 9a ,b with saturated metha-
nolic ammonia afforded in good yield the free nucleosides
11a ,b and 12a ,b, respectively.
a
Reagents and conditions: (a) LiCl, DMSO, 200 °C (bath
temperature), 2 h (2, 55%; 5, 72%); (b) H2, 5% Pd/C, EtOH, rt, 18
h; (c) BzCl, pyridine, rt,18 h (b + c, sum 71%); (d) persilylated
base, (1) 5/thymine/(TMS)Tfl, 1.0:2.0:3.0, mol; CH3CN, reflux, 5 h
(5 + 6, 19%; 7a ,b, 49%); (2) 5/N6Bz-adenine/SnCl4, 1.0:2.0:5.0, mol;
CH3CN/1,2-dichloroethane (2:1, vol), reflux, 5 h (5 + 6, 7.5%; 8a ,
9%, 8b, 66%); (3) 5/N4Bz-cytosine/SnCl4, 1.0:2.0:3.0, mol; CH3CN,
reflux, 5 h (5 + 6, 18%; 9a , 8%; 9b, 56%); (e) NH3/MeOH, rt, 24 h
(10a , 28%; 10b, 55%; 11a , 78%; 11b, 83%; 12a , 80%; 12b, 78%).
The radical-mediated hydrodehalogenation of a mix-
ture of 7a ,b with tributyltin hydride in anhydrous
dioxane in the presence of R,R′-azobisisobutyronitrile
(AIBN) and subsequent debenzoylation and chromatog-
raphy gave FLT (13b) and its R-anomer 13a (RFLT) in
55% and 27% yield, respectively (Scheme 2). Similar
transformation of deprotected â-anomer 10b afforded
FLT in 90% yield. Adenine nucleosides 8a and 8b were
easily converted in excellent yields into the corresponding
R- and â-fluorides 14a (RFLA) and 14b (FLA), whereas
the dehydrohalogenation of cytosine nucleoside 9b gave
a complex reaction mixture, from which nucleoside 15b
(FLC) was isolated in 31% yield after debenzoylation and
chromatography. In contrast to this, treatment of depro-
tected cytosine nucleosides 12a and 12b with tributyltin
hydride in 1:1 dioxane-methanol gave the corresponding
R- and â-fluorides 15a (RFLC) and 15b (FLC) in high
yield, although the reaction requires more prolonged
heating. In a similar way, nucleosides 10b and 11b were
converted into FLT and FLA, respectively.
trile/1,2-dichloroethane (2:1, v/v) mixture for 5 h followed
by column chromatography gave 1-(5-O-benzoyl-2-chloro-
2,3-dideoxy-3-fluoro-â-D-ribofuranosyl)-N6-benzoylade-
nine (8b; 66%) and its R-anomer 8a (9%) as well as a
mixture of the methyl glycosides 5 and 6 (7.5%). An
excess of SnCl4 was employed to minimize the formation
of the N7-adenine nucleoside(s).35-37 No evidence for the
presence of the N7-isomer(s) was detectable by means of
TLC. It is noteworthy that the reaction of the furanoside
5 with silylated N6-benzoyladenine proceeded with better
stereoselectivity (a ca. 1:7 R:â ratio) of the glycosidic bond
formation compared to the similar reaction of methyl
2-azido-5-O-benzoyl-2,3-dideoxy-3-fluoro-â-D-ribofurano-
side (a ca. 1:3 R:â ratio).
The reaction of silylated N4-benzoylcytosine with the
furanoside 5 in the presence of excess SnCl4 (2.0:1.0:3.0,
mol) in refluxing acetonitrile for 5 h gave, after standard
workup and column chromatography, 1-(5-O-benzoyl-2-
chloro-2,3-dideoxy-3-fluoro-â-D-ribofuranosyl)-N4-benzoyl-
cytosine (9b; 56%), its R-anomer 9a (8%), and a mixture
of the methyl glycosides 5 and 6 (18%). Thus, this
reaction gave better results in terms of both the yields
of the nucleosides (ca. 64%, combined) and the R:â ratio
In extension of this work, we studied the dehydroha-
logenation of nucleosides 8a ,b, 10b, and 11b under the
action of 1 M MeONa in methanol (Scheme 3). Treatment
of nucleoside 10b with 1 M MeONa/MeOH under reflux
for 5 h followed by silica gel column chromatography gave
the chloroolefin 16 (yield 46%) and unchanged starting
compound (17%).39 The preferred elimination of HF
instead of HCl is unexpected on the basis of the greater
strength of the C-F bond (105.5 kcal/mol) compared to
(36) Akhrem, A. A.; Adarich, E. K.; Kulinkovich, L. N.; Mikhailopulo,
I. A.; Poschastieva, E. B.; Timoshchuk, V. A. Dokl. Acad. Nauk SSSR
1974, 219, 99-101; Chem. Abstr. 1975, 82, 86532.
(37) Poopeiko, N. E.; Kvasyuk, E. I.; Mikhailopulo, I. A.; Lidaks, M.
J . Synthesis 1985, 605-608.
J . Org. Chem, Vol. 68, No. 15, 2003 5899

Mikhailopulo et al.
SCHEME 3a
These considerations imply that the HF elimination
leading to the olefins proceeds via a concerted trans
elimination by the attack of the MeO- anion at the C2′
proton as a driving force.
In contrast to the above-discussed HF elimination,
methyl 3-deoxy-3-fluoro-2-O-mesyl-â-D-ribofuranoside was
transformed to the vinyl 3-fluoride on treatment with
t-BuOK.41 Similarly, 2′-deoxy-2′-fluoro-3′-O-mesyl-â-D-
ribofuranosyl nucleosides were converted to vinyl 2-flu-
orides under the action of t-BuOK41 or NaOH in ethanol.42
The stereochemistry of these fluoromesylates was not
investigated to the best of our knowledge. Marquez and
co-workers suggested, however, the formation of the vinyl
fluorides via trans elimination, implying a major popula-
tion of the S-type conformation of the furanose ring of
the starting N6-benzoyl-2′-deoxy-5′-O-dimethoxytrityl-2′-
fluoro-3′-O-mesyladenosine.42
NMR Sp ectr oscop ic Stu d ies. The structure of the
synthesized compounds is corroborated by 1H (Tables S1
and S2 in the Supporting Information) and 13C (Tables
S3 and S4 in the Supporting Information) NMR data and
by UV spectroscopy (see the Experimental Section). The
most informative features of the NMR spectra of the
R-anomers are (i) the 0.26-0.60 and 0.25-0.39 ppm shifts
of the respective H1′ and H4′ resonance signals to a lower
field on going from the â-anomers to R-anomers and (ii)
the long-range couplings between the fluorine atom and
C6 of the thymine base of 10a (5J 6,F ) 7.85 Hz) and C8
of the adenine base of 11a (5J 8,F ) 11.63 Hz) in the 13C
NMR spectra. The latter couplings are generally indica-
tive of a spatial proximity of the nuclei involved20,43,44 and
are not observed in the corresponding â-anomers 10b and
11b.
a
Reagents and conditions: (a) 1 M MeONa in MeOH, reflux
for (i) 5 h (16, 46%; 18, 47%), (ii) 4 h (19, 82% from 8b and 77%
from 11b; 21, 69%), and (iii) 1 h (22, 86%).
that of the C-Cl bond (78.5 kcal/mol38). A similar trend
is displayed by 1-(2-azido-2,3-dideoxy-3-fluoro-â-D-ribo-
furanosyl)thymine (17):35 treatment of 17 with 1 M
MeONa/MeOH under reflux for 5 h furnished, after
standard workup and subsequent chromatography, the
vinyl azide 1840 as the main product, along with the
unconsumed starting compound 17 in 47% and 33%
isolated yield, respectively.
Adenine nucleosides 8a ,b, 11b, and 2035 displayed a
similar course of HF elimination on treatment with 1 M
MeONa/MeOH under reflux, affording the corresponding
vinyl chlorides 19 and 22 and the vinyl azide 21 (Scheme
3). It is noteworthy that the rate of conversion of the
R-nucleoside 8a to the R-vinyl chloride 22 was much
higher than that of the analogous reaction with the
â-anomers 8b and 11b.
A likely explanation of the HF vs HCl elimination is
suggested by the conformational peculiarities of the
starting nucleosides (vide supra). As can be seen from
Table 1, the strong predominance of the south (S) type
conformer of the furanose ring is a common characteristic
of these nucleosides. It should be stressed that the phase
angle (P) of the pseudorotational equilibrium is in a
narrow domain (150° e P e 162°) characteristic of the
2E furanose puckering. In this conformation of the
D-ribofuranose ring a fluorine atom and the C2′ proton
are trans-oriented (antiperiplanar, ap) whereas a chlorine
atom and the C1′ and C3′ protons are gauche-arranged
(synclinal, sc) in both the â- and R-nucleosides studied.
1
The H NMR data of the fluorides 13-15 are in accord
with those measured for the same compounds prepared
by alternative procedures.12,18,24 The structures of the
1
olefins 16, 18, 19, 21, and 22 were confirmed by H and
13C NMR spectroscopy and are in agreement with the
reported data for closely related compounds.39,40,45
A qualitative inspection of the absolute values of 3J FC1′
3
and J FC5′ (Table S4) of the nucleosides 10a ,b, 11a ,b, 12b,
17, and 20 point clearly to a high preference for the
S-type pentofuranose ring conformation. In this confor-
mation, the F-C3′-C4′-C5′ and F-C3′-C2′-C1′ frag-
ments are in antiperiplanar (ca. 170°) and gauche (ca.
90°) arrangements, respectively, which is consistent46
with the corresponding 3J CF values within the range 9.1-
11.6 Hz and of <2.0 Hz (Table S4). It is striking that the
â f R change of the anomeric configuration does not seem
to give rise to an essential displacement of the predomi-
nant population of the S-type puckered conformers (see
below). By contrast, sugars 2-6 show a conformational
3
3
diversity as can be seen from both the J HH and J FH
couplings (Table S2).
(41) Nakayama, T.; Asai, T.; Okazoe, T.; Suga, A.; Morizawa, Y.;
Yasuda, A. Nucleic Acids Res. 1991, Special Issue No. 18, 191-192.
(42) Siddiqui, M. A.; Driscoll, J . S.; Marquez, V. E. Tetrahedron Lett.
1998, 39, 1657-1660.
(38) Ingold, C. K. Structure and Mechanism in Organic Chemistry,
2nd ed.; Cornell University Press: Ithaca, NY, and London, 1969;
Chapter III.
(39) Van Aerschot, A.; Herdewijn, P. Bull. Soc. Chim. Belg. 1989,
98, 931-936.
(40) Mikhailopulo, I. A.; Zaitseva, G. V.; Vaaks, E. V.; Balzarini, J .;
De Clercq, E.; Rosemeyer, H.; Seela, F. Liebigs Ann. Chem. 1993, 513-
519.
(43) Mele, A.; Salani, G.; Viani, F.; Bravo, P. Magn. Reson. Chem.
1997, 35, 168-174.
(44) Mele, A.; Vergani, B.; Viani, F.; Meille, S. V.; Farina, A.; Bravo,
P. Eur. J . Org. Chem. 1999, 187-196.
(45) J ain, T. C.; J enkins, I. D.; Russel, A. F.; Verheyden, J . P. H.;
Moffatt, J . G. J . Org. Chem. 1974, 39, 30-38.
(46) Kalinowski, H.-O.; Berger, S.; Braun, S. 13C NMR Spectroscopie;
G. Thieme Verlag: Stuttgart and New York, 1984; pp 526-529.
5900 J . Org. Chem., Vol. 68, No. 15, 2003

2′-Chloro-2′,3′-dideoxy-3′-fluoro-D-ribonucleosides
TABLE 1. P seu d or ota tion a l P a r a m eter s of Som e 2′,3′-Did eoxy-3′-flu or o-2′-(X-su bstitu ted )-D-r ibon u cleosid es
PN
(deg)
ψN
(deg)
PS
(deg)
ψS
(deg)
rms
(Hz)
∆Geffa(2′-3′)
∆Geffa(3′-4′)
∆J
(Hz)
∆GS/N
compd
base
C2′-X
(Hz)
(Hz)
% S
(kJ ‚mol-1
)
â-Series
0.01
0.02
0.01
0.01
0.02
0.4
10bb
10bc
11bb
11bc
12bc
13bc
14bb
F LGb
17b
T
T
Cl
Cl
Cl
Cl
Cl
H
H
H
N3
N3
N3
NH2
NH2
NH2
OH
F
18e
18e
18e
18e
18e
-20
0
36e
36e
36e
36e
36e
32e
32e
32e
33e
40e
38e
36e
36e
36e
33e
37.6
155.5
157.0
161.8
164.4
165.0
160.6
171.9
167.9
152.3
154.6
155.7
150.1
159.5
159.3
175.3
165.4
36.9
36.4
35.4
35.1
35.0
32.2
31.6
32.0
32.4
36.7
36.8
39.3
36.7
36.7
32.0
37.0e
-6.7
-7.1
-6.7
-5.6
-9.1
4.7
4.1
4.1
-5.8
-5.6
6.7
6.3
7.6
7.1
6.3
7.1
8.2
8.2
6.0
6.3
6.3
6.0
8.4
6.9
3.9
7.4
0.0
0.0
0.0
0.0
0.0
0.6
0.3
0.3
0.0
0.0
0.1
0.0
0.0
0.0
0.0
0.0
97
95
95
94
95
94
98
96
99
94
95
97
95
95
98
80
-9.1
-7.7
-7.7
-7.2
-7.7
-7.2
-9.1
-8.3
-9.1
-7.2
-7.7
-9.1
-7.7
-7.7
-9.1
-3.6
A
A
C
T
A
G
T
0.2
0.2
0
18e
18e
18e
18e
18e
18e
18e
13.2
0.01
0.01
0.03
0.01
0.01
0.01
0.1
17c
T
20c
A
A
T
C
A
A
23b
-2.2
-2.6
-3.0
-2.6
-8.2
24c
25b
26d,f
27c,g
0.0
R-Series
0.1
10a c
11a c
12a c
13a c
14a b
rF LGb
T
Cl
Cl
Cl
H
H
H
58.3
0.3
-7.8
36e
0e
33e
33e
33e
36e
36e
36e
185.7
168.1
189.6
167.3
173.3
180.3
31.6
30.8
40.1
29.2
27.6
26.7
-3.5
-3.5
-3.5
7.3
6.0
6.0
3.7
3.7
3.7
4.1
5.4
5.4
0.1
0.3
0.2
0.1
0.3
0.3
82
94
58
97
95
96
-4.0
-7.2
-0.9
-9.1
-7.7
-8.3
A
C
T
A
G
0.1
0.1
0.1
0.2
0e
0.1
a
b
d
∆Geff ) -3.72[(RFCC + RHCC)/2 - 110] along a given C-C bond (see the text). DMSO. c CD3OD. CDCl3. e Kept fixed during the final
minimization. f 5′-O-Benzyl derivative.63 See ref 64.
g
Con for m a tion a l An a lysis
fluorine coupling constants and the corresponding H-C-
C-F torsion angles including correction terms for
An a lysis of P seu d or ota tion P a r a m eter s. The con-
formational analysis of the furanose rings of selected
sugars and nucleosides described was performed with the
aid of the PSEUROT (version 6.3) program, which
substituent electronegativity and for the H-C-C (RHCC
)
and F-C-C (RFCC) bond angle changes from the tetra-
hedral values.48 The latest release (version 6.3) of the
PSEUROT program incorporates the published Karplus
3
3
calculates the best fits of three J HH and two J HF
3
parameters for J HF and allows simultaneous analysis of
experimental coupling constants [3J H1′H2′ 3J H2′H3′, and
,
3
3J HH and J HF experimental coupling constants, which
3
3
3J H3′H4′, and J H2′F3′ and J H4′F3′] to the five conformational
parameters (phase angles P and puckering amplitudes
ψm for both the north (N) and the south (S) type
conformers and the corresponding molar ratio).47 In the
PSEUROT program a minimization of the differences
between the experimental and calculated couplings is
accomplished by a nonlinear Newton-Raphson minimi-
zation; the quality of the fit is expressed by the root-
mean-square (rms) difference. This procedure presup-
poses the existence of a two-state N/S equilibrium. When
only three vicinal 3J HH coupling constants along the ring
bonds exist, as is the case in the present series, two of
the five conformational parameters must be kept fixed
to preset values during the minimization. This fixing is
usually not much of a problem whenever the equilibrium
is strongly one-sided (major form g85%) but may lead to
more-or-less wide ranges of possible conformational
parameters otherwise. These ranges can be narrowed by
increasing the ratio of the number of data points vs the
number of optimized parameters. To achieve this end,
two different strategies can be employed: (i) measure-
ment of the NMR spectra over a wide range of temper-
atures; (ii) the inclusion of other coupling information in
the calculation.
greatly facilitates the conformational analysis of pento-
furanose rings because of the increase in the number of
experimental data points over the puckering parameters
P and ψm.47,48 We propose to write the bond angle
correction term as
∆3J HF ) ∆Geff cos2 ΦHF
(1)
where ∆Geff ) G[(RFCC + RHCC)/2 - 110] (G ) -3.72 Hz)
and ΦHF is the H-C-C-F torsion angle under consid-
eration.
In the original work48 the bond angles R were deter-
mined from 3-21G ab initio calculations and the param-
eter G was optimized, together with the remaining
Karplus parameters A-F, to fit both J HH and J HF
simultaneously. Lacking ab initio data spanning pseu-
dorotational space for the compounds in the present
collection, we chose to invert the procedure and fitted
∆Geff to minimize ∆J HF along each bond (Tables 1 and
2). Note that this approach involves, first of all, the
determination of the best pseudorotation parameters or
3
range of best parameters on the basis of J HH only and
then fixing the conformers found to manually fit ∆Geff
along the coupling paths in question. Another main
criterion used here was that ∆Geff in a series of similar
compounds should be fairly constant in sign and magni-
tude. A rough correlation with the electronegativity of
Chattopadhyaya and co-workers have recently devel-
oped a new Karplus-type relation between vicinal proton-
(47) van Wijk, J .; Haasnoot, C. A. G.; de Leeuw, F. A. A. M.;
Huckriede, B. D.; Westra Hoekzema, A. J . A.; Altona, C. PSEUROT
6.3; Leiden Institute of Chemistry, Leiden University: Leiden, The
Netherlands, 1999.
(48) Thibaudeau, C.; Plavec, J .; Chattopadhyaya, J . J . Org. Chem.
1998, 63, 4967-4984.
J . Org. Chem, Vol. 68, No. 15, 2003 5901

Mikhailopulo et al.
TABLE 2. P seu d or ota tion a l P a r a m eter s of Som e Meth yl 2,3-Did eoxy-3-flu or o-2-(X-su bstitu ted )-D-r ibofu r a n osid es
PN
(deg)
ψN
(deg)
PS
(deg)
ψS
(deg)
rms
(Hz)
∆Geffa(2′-3′)
∆Geffa(3′-4′)
∆J
(Hz)
∆GS/N
compd
5-O-R
C2′-X
(Hz)
(Hz)
% S
(kJ ‚mol-1
)
â-Seriesb
0.0
2
4
5
Bn
H
Bz
Bn
H
Cl
Cl
Cl
F
F
H
H
-13.8
-7.5
-22.5
0.1
10.4
-59.8
-67.3
42.1
31.9
38.4
34.9
32.4
42c
166.0
179.5
167.2
171c
39c
39c
38c
-8.2
-8.2
-6.7
-10.8
-10.8
0.4
3.5
3.5
6.3
2.1
2.1
0.0
0.0
0.0
0.2
0.2
0.2
0.4
41
40
10
15
18
69
64
0.9
1.1
5.7
4.5
4.0
-2.1
-1.5
0.0
0.0
28d
29d
30d
31d
36.3
32.4
40.1
43.5
0.1
0.1
0.1
0.2
153c
Bn
H
225.5
218.8
10.2
10.2
42c
0.4
R-Seriesb
0.0
6I
Bz
Bz
Bz
Cl
Cl
OH
-55.5
-61.1
-47.0
33.5
37.5
28c
142.4
137.3
145.2
36.9
36.5
28.0
6.7
2.1
0.2
2.1
2.1
-0.7
0.2
0.0
0.0
63e
76e
96
-1.4
-3.0
-8.3
6II
0.0
0.0
32f
a
b
∆Geff ) -3.72[(RFCC + RHCC)/2 - 110] along a given C-C bond (see the text). All 1H NMR spectra have been measured in CDCl3.
c Kept fixed during the final minimization. See ref 64. e These examples serve to illustrate the unusual uncertainty in the conformational
d
populations of 6. Actually, all populations between about 60% S and 90% S are compatible with the data. f See ref 63.
at C1′ means a drive toward the N form in the â-D-series
and toward the S form in the R-D-series. Interestingly, it
was demonstrated experimentally that the strength of
this drive can be manipulated by protonation or depro-
tonation of the nucleobase.50,51 Following convention, the
associated enthalpy and free-energy changes are termed
positive when the drive favors the N-type conformers and
negative when it favors the S-type conformers.
In the discussion of the AE presented here the so-called
F IGURE 1. Schematic presentation of the north (N; C2′-exo/
C3′-endo) h south (S; C2′-endo/C3′-exo) pseudorotational equi-
Altona-Havinga model49 will be used, which entails n
f σ* stabilization, where n represents an available lone
pair of electrons and σ* is a suitably positioned anti-
bonding orbital. It is important to note that quantum-
chemical calculations of gauche and trans conformers in
XCH2OH indicate a linear correlation between the
librium in â-D-nucleosides (R1 ) base) and -pentofuranose
sugars (R1 ) OR, where R ) H, Me, etc).
neighboring substituents is expected and found (Tables
1 and 2). The nature of the solvent also seems to play a
role.
strength of the AE and the electronegativity of X (øCH
<
3
ø
NH2
< øOH < øF).52 In practice, the matters are often more
The necessary A and B parameters, which relate the
endocyclic to the exocyclic torsion angles between cou-
pling nuclei,47 were taken from the literature48 and
averaged for similar structures. To account for the facts
that the magnitude of 3J HF is roughly 5 times larger than
the magnitude of 3J HH and that the Karplus parameters
for the HF coupling were obtained on a relatively small
set of experimental J values, scale factors were intro-
complicated by competition among various stereoelec-
tronic factors at the same anomeric center,53 and the
anomeric strength of a given aglycon varies with the
chemical nature of nearby substituents.
The term “GE” was coined by Wolfe,54a who showed
that a vicinal fragment such as X-C-C-Y prefers to
adopt a gauche conformation along the central C-C bond
in cases where X and Y represent highly electronegative
ligands (or electron pairs) although dipole-dipole repul-
sions and steric factors work in favor of the antiperipla-
nar form. In work with the aid of the earliest version of
the PSEUROT program Haasnoot et al.54b suggested that
the conformational behavior of a series of nucleic acid
constituents could be rationalized on the basis of the GE
concept. The possible origin(s) of the GE merits some
discussion here because different views may lead to
different conclusions.
3
3
duced (1.0 for J HH and 0.1 for J HF).
Ster eoelectr on ic F a ctor s. Besides steric interac-
tions, the north (N) (C2′-exo/C3′-endo)/south (S) (C2′-endo/
C3′-exo) pseudorotational equilibrium in nucleosides,
nucleotides, and pentafuranose sugars (Figure 1) is
driven by competing stereoelectronic factors known as
the anomeric effect (AE) and gauche effect (GE).
A broad definition of the AE is embodied in the
statement that, in a fragment C-X-C-Y where ligand
X carries one or two lone pairs of electrons and Y
represents an electronegative ligand, the gauche con-
formers along the central X-C bond are preferred over
the trans form (antiperiplanar, ap).49 In nucleosides and
related sugars this means that an electronegative ligand
L at C1′ is driven to adopt the pseudoaxial position
(gauche C4′-O4′-C1′-L). Of course, steric forces coun-
teract this drive. Preference for a pseudoaxial position
The first view focuses on the gauche situation itself
(50) Thibaudeau, C.; Fo¨ldesi, A.; Chattopadhyaya, J . Tetrahedron
1998, 54, 1867-1900.
(51) Thibaudeau, C.; Plavec, J .; Chattopadhyaya, J . J . Org. Chem.
1996, 61, 266-286.
(52) Smits, G. F.; Krol, M.; Altona, C. Mol. Phys. 1988, 65, 513-
529.
(53) Krol, M. C.; Huige, C. J . M.; Altona, C. J . Comput. Chem. 1990,
11, 765-790.
(49) (a) The Anomeric Effect and its Associated Stereoelectronic
Effects; Thatcher, G. R. J ., Ed.; ACS Symposium Series; American
Chemical Society: Washington, DC, 1993. (b) Kirby, A. J . The Anomeric
Effect and Related Stereoelectronic Effects at Oxygen; Springer-Ver-
lag: Berlin, 1983.
(54) (a) Wolfe, S. Acc. Chem. Res. 1972, 5, 102-111. (b) Haasnoot,
C. A. G.; de Leeuw, F. A. A. M.; de Leeuw, H. P. M.; Altona, C. Org.
Magn. Reson. 1981, 15, 43-52. (c) Plavec, J .; Tong, W.; Chatto-
padhyaya, J . J . Am. Chem. Soc. 1993, 115, 9734-9746. (d) Thibaudeau,
C.; Chattopadhyaya, J . Nucleosides Nucleotides 1997, 16, 523-529.
5902 J . Org. Chem., Vol. 68, No. 15, 2003

2′-Chloro-2′,3′-dideoxy-3′-fluoro-D-ribonucleosides
and accepts either an attraction between the gauche-
oriented electronegative ligands X and Y,54a which from
this point on will be denoted as the standard view as it
is most widely used, or a relative destabilization of
antiperiplanar X-C-C-Y because of bond bending.55 For
example, it is often stated54c,d that in â-D-ribonucleosides
the GE of O4′-C1′-C2′-O2′ favors a pseudoaxial O2′
ligand, i.e., drives the ribonucleoside toward the N form,
whereas this drive is counteracted by the GE of N-C1′-
C2′-O2′, which will drive both the nucleobase and O2′
to adopt pseudoequatorial positions, i.e., the S form. In
a similar vein, the GE of O4′-C4′-C3′-O3′ in â-D-
ribonucleosides drives them toward the S form with the
O3′ ligand in a pseudoaxial position. Finally, in this view
the fragment X2′-C2′-C3′-Y3′ remains gauche in either
the N or S form and the GEs of both forms effectively
cancel each other. However, a closer look at the geom-
etries involved in the first three of the above examples,
which involve interactions between an exocyclic and an
endocyclic ligand, shows that in the Newman projections
the relevant bonds are approximately perpendicular to
one another (90° ( 12°). This means that any orbital-
orbital interaction will tend to be minimal.
on the electronegativity of the ligands themselves. More-
over, we assume that APs between like ligands, for
example, AP[OR-OR], do not contribute to the confor-
mational energy. At this point three types of APs can be
distinguished: (i) neutral APs, which presumably do not
influence the drive toward one conformer or another, for
example, AP[H3′-H4′] in the N-type ribosides; (ii) con-
stant APs, which are in the first approximation only
dependent on the electronegativity difference between the
two ligands at either end of the intervening bonds, for
example, AP[O4′-H3′] favoring the S form and AP[O4′-
H2′] favoring the N form of ribose rings; in practice, these
APs are not constant but modulated by other substituents
present; (iii) potentially variable APs, the driving force
of which is expected to increase with the electronegativity
difference between the ligands L and L′ at each end of
the central bond. Moreover, the relatively small differ-
ences between the various AP torsion angles in ribosides
(usually ranging from 140° to 160°) are neglected. The
following dissection is based on hypothetical â- and R-D-
2,3-dideoxy-X2,Y3-ribosides; the aglycon is indicated by
R1. For obvious reasons the neutral APs such as AP[H-
H] and AP[C-H] need no discussion.
A radically different view on the GE, proposed by
Brunck and Weinhold,57a is embodied in the statement
that this effect arises from a relative stabilization of a
situation where the most electronegative ligand is ap-
proximately antiperiplanar to the least electronegative
ligand. In analogy with the antiparallel nX T σ*C-Y lone
pair-antibonding donor-acceptor orbital mixing expla-
nation of the AE,56 Brunck and Weinhold suggest maxi-
mum antiperiplanar σ T σ* stabilization where the
donating bond is the least polar one (usually but not
necessarily C-H) and the acceptor orbital is at the most
polar bond.57a A recent quantum-chemical study57b sup-
ports this model. According to the second view the term
“gauche effect” is confusing57a and could be more aptly
named the antiperiplanar effect (AP). To avoid misun-
derstandings, the accepted name is best retained. How-
ever, for the sake of clarity it is desirable at this point to
introduce a modified notation, which does not refer to
the gauche orientation of the polar bonds but instead
draws attention to the AP situations (favorable and
unfavorable), which are available to the fragment con-
sidered. Note that lone electron pairs are not counted as
ligands here. As our examples pertain to the furanose
rings shown in Tables 1 and 2, the standard notation is
often also given to clarify the differences between the two
views discussed. It is better to keep in mind that along
each C-C bond in a nonplanar ring and for a given
conformer there is one AP interaction between a vicinal
pseudodiaxial (aa) exocyclic ligand pair against two AP
interactions between each pseudoequatorial exocyclic
ligand and a ring atom. In the case of a two-conformer
equilibrium, one has to examine six APs per bond. As
the central bond is always C-C there is no need to refer
explicitly to the so-called L-polar bonds57a at each end of
the L-C-C-L′ fragment, and it suffices to concentrate
Let us first look along the C3′-C4′ and C2′-C3′ bonds
of ribosides, concentrating on the APs involving Y3′. In
the N form one has two variable APs: (i) AP[O4′-Y3′],
the energy of which will vary from highly stabilizing in
the case of Y3′ ) H to zero for Y3′ ) OR and possibly
even destabilizing upon a further increase of the elec-
tronegativity ø of Y3′; (ii) AP[Y3′-C1′], which, however,
balances (cancels) the variable AP[Y3′-C5′] of the S form.
In the S form, one has additionally the variable and
stabilizing AP[Y3′-H2′] and the constant and stabilizing
AP[O4′-H3′]. The well-known GE of O4′-C4′-C3′-Y3′
thus represents the sum of six APs; however, the vari-
ability of this GE with øY rests solely upon the balance
between AP[O4′-Y3′] and AP[Y3′-H2′], and the net
effect will be an increasing stabilization of the S form
relative to the N form upon increasing øY3′. In fact, it was
experimentally demonstrated58 that in a series of 3′-
substituted 2′,3′-dideoxythymidines the enthalpy ∆H°GE
linearly correlates with øY3′ on Mullay’s scale.59 It was
also shown60 that in the abasic sugar skeleton the GE of
O4′-C4′-C3′-O3′ and the GE of O4′-C1′-C2′-O2′ are
of equal magnitude and of opposite sign. In terms of APs
this magnitude is equal to two AP[O-H] interactions if
one neglects the small difference between AP[O2′-H1′]
and AP[O3′-C5′]. Other work61 has revealed that the
nominal strength of the GE of O4′-C4′-C3′-O3′ in â-D-
2′-deoxynucleosides appears to depend on the nature of
the nucleobase.
Along the C2′-C3′ bond the overall GE of X2′-C2′-
C3′-Y3′ may slightly differ according to the viewpoint
taken. The first view predicts that, as the geometry
remains gauche in either the N or S form, there will not
be any detectable drive toward one side or another
regardless of the electronegativities of X2′ and Y3′.
(58) Thibaudeau, C.; Plavec, J .; Garg, N.; Papchikin, A.; Chatto-
padhyaya, J . J . Am. Chem. Soc. 1994, 116, 4038-4043.
(59) (a) Mullay, J . J . Am. Chem. Soc. 1985, 107, 7271-7275. (b)
Mullay, J . J . Am. Chem. Soc. 1984, 106, 5842-5847.
(60) Luyten, I.; Thibaudeau, C.; Chattopadhyahya, J . J . Org. Chem.
1997, 62, 8800-8808.
(55) Wiberg, K. B. Acc. Chem. Res. 1996, 29, 229-234.
(56) Romers, C.; Altona, C.; Buys, H. R.; Havinga, E. Top. Stereo-
chem. 1969, 4, 39-97. Altona, C. Thesis, Leiden University, Leiden,
The Netherlands, 1964.
(57) (a) Brunck, T. K.; Weinhold, F. J . J . Am. Chem. Soc. 1979, 101,
1700-1709. (b) Rablen, P. R.; Hoffmann, R. W.; Hrovat, D. A.; Borden,
W. T. J . Chem. Soc., Perkin Trans. 2 1999, 1719-1726.
(61) Plavec, J .; Thibaudeau, C.; Chattopadhyaya, J . Pure Appl.
Chem. 1996, 68, 2137-2144.
J . Org. Chem, Vol. 68, No. 15, 2003 5903

Mikhailopulo et al.
SCHEME 4
According to the second view, one distinguishes on one
hand two variable APs that favor the N form, the
pseudodiaxial AP[X2′-H3′] and the endocyclic AP[Y3′-
C1′], vs two APs that favor the S form, the pseudodiaxial
AP[Y3′-H2′] and the endocyclic AP[X2′-C4′]. On first
sight, one expects virtually complete mutual cancellation
of these APs and thus a net drive of zero because in the
normal N and S pseudorotational ranges the respective
AP torsion angles are approximately constant, amounting
to roughly (155°. A small differential effect could occur
when the electronegativity of C1′ is modified inductively
by a change of aglycon.
Along the C1′-C2′ bond two GEs are customarily
recognized, the GE of O4′-C1′-C2′-X2′ and the GE of
R1′-C1′-C2′-X2′, the former driving toward the N form
in cooperation with the anomeric effect in the â-D-
anomers and counteracted by the AE in the R-D-anomers
and the latter driving toward the S form in â-D-anomers
but balancing in the R-D-anomers. In terms of APs,
matters look more complicated. Considering the â-D-
ribonucleosides first (R1′ ) B), one recognizes that the
N form, besides the favorable anomeric effect, is charac-
terized by a strongly favorable but constant AP[O4′-H2′]
and by a highly variable pseudodiaxial AP[B-X2′].
Unfortunately, the electronegativity øB of any nucleobase
on Mullay’s scale is unknown, but it is safe to assume øC
< øB < øO. In the case of 2′-deoxynucleosides (X2′ ) H2′)
the AP[B-H2′] reinforces the AE drive toward the N
form, but for X2′ ) Cl, O, or F little can be said for the
moment. In the S form, one has the variable AP[O4′-
X2′] and a base-dependent variable AP[B-C3′]; in the
â-D-2′-deoxy case the former AP balances AP[O4′-H2′].
In â-D-ribonucleosides AP[O4′-O2′] is zero. It follows that
the net effect of base changes on AP energies along the
C1′-C2′ bond is probably relatively small compared to
the AE. In the N form of R-D-ribonucleosides one again
encounters the constant N-driving AP[O4′-H2′], coun-
teracted by the S-driving AP[O4′-H3′], as well as the
strongly variable AP[X2′-H1′], the N drive of which runs
from zero for X2′ ) H to a maximum value for X2′ ) F.
The S form is mainly characterized by the variable AP-
[X2′-O4′], which runs from highly stabilizing for X2′ )
H (balancing AP[O4′-H2′] in the N form) to zero for X2′
) O, and by the variable AP[B-H2′] augmenting the AE.
Of course, the AE is now S-driving. In methylated sugars
(R1′ ) OMe), more specific predictions can be made, and
these will be treated in the next section.
and 2′,3′-dideoxy-2′,3′-difluoroadenosine (27),64 are in-
cluded in Table 1 (Scheme 4). Note that the pseudorota-
tional parameters for 9-(2-amino-2,3-dideoxy-3-fluoro-â-
D-ribofuranosyl)guanine are identical with those of cytosine
nucleoside 25 and are, therefore, not included in Table
1. The carbohydrate precursors in the synthesis of 27,
viz., methyl 2,3-dideoxy-2,3-difluoro-5-O-benzyl-â-D-ri-
bofuranoside (28) and its 5-O-debenzoylated derivative
29,64 are included in Table 2. Moreover, methyl 2,3-
dideoxy-3-fluoro-5-O-benzyl-â-D-erythro-pentofurano-
side (30) and its derivative 3164 are also included in Table
2.
Due to the fact that only free-energy differences are
available from the present work the following analysis
is necessarily of a qualitative nature. The 3′-deoxy-3′-
fluoro-â-D-ribonucleoside series (substituent X2′ ) H, Cl,
N3, NH2, OH, and F) shows remarkable constant behav-
ior. Except for X2′ ) F, all compounds display a strong
preference (g94%) for the S-type sugar conformer re-
gardless of the nature of the nucleobase and regardless
of the electronegativity of the X2′ substituent. This means
that the pseudoaxial F3′ ligand of the S-form completely
dominates the situation. Disregarding the APs that
mutually cancel in the S vs N form, for example, AP-
[O4′-H3′] vs AP[O4′-H2′], the strong S drive must be
due to AP[F3′-H2′], far outweighing the anomeric effect
of the nucleobase. Only for X2′ ) F2′ this dominance is
partly broken (80% S). In the first approximation one
could expect that AP[F2′-H3′] in the N form would
perfectly match AP[F3′-H2′] in the S form. Nevertheless,
the strongly N-driving AE of the base does not gain the
upper hand. A closer look reveals that the S vs N AP
situation is not symmetrical because the former has AP-
[F3′-C5′], which is expected to be strongly S-driving
compared to the relatively neutral AP[F2′-B] in the N
form. In addition, it is likely that the anomeric effect of
the base diminishes with increasing electronegativity of
X2′, judging by the conclusion that the anomeric effect
of Ade in the â-D-ribonucleoside is smaller than the AE
in its 2′-deoxy counterpart.51 In terms of the standard
view the rationalization would involve the balancing of
all GEs except for a strongly S-driving GE(F2′-C2′-C1′-
B) counteracting the N-driving AE.
P SEUROT Ca lcu la tion s. The conformational data on
2′,3′-dideoxy-3′-fluoro-2′-(X-substituted)-â,R-D-ribonucleo-
sides are collected in Table 1 and those on some similarly
substituted methyl glycosides in Table 2. To gain more
detailed insight into the role of a ribo substituent at C2′,
the PSEUROT analyses of 9-(2,3-dideoxy-3-fluoro-â-D-
erythro-pentofuranosyl)guanine (FLG) and its R-anomer
(RFLG),24 2′-amino-2′,3′-dideoxy-3′-fluoro-â-D-ribonucleo-
sides of adenine (23), thymine (24), and cytosine (25),62
as well as 5′-O-benzyl-3′-deoxy-3′-fluoroadenosine (26)63
In the R-D-nucleoside series (Table 1), only two differ-
ent X2′ substituents are at hand, H and Cl. For X2′ ) H
(62) Mikhailopulo, I. A.; Sivets, G. G.; Pricota, T. I.; Poopeiko, N.
E.; Balzarini, J .; De Clercq, E. Nucleosides Nucleotides 1991, 10, 1743-
1757.
(63) Mikhailopulo, I. A.; Sivets, G. G. Helv. Chim. Acta 1999, 82,
2052-2065.
(64) The synthesis and characterization are described in the Sup-
porting Information.
5904 J . Org. Chem., Vol. 68, No. 15, 2003

2′-Chloro-2′,3′-dideoxy-3′-fluoro-D-ribonucleosides
(13a , 14a , RFLG) either view would predict (in different
words) a strong predominance of the S form, and this is
found (g95% S). Matters are different for X2′ ) Cl (10a ,
11a , 12a ). In going from Ade (11a , 94% S) to Thy (10a ,
82%S) to Cyt (12a , 58%S) a remarkable base-dependent
progression in the S/N population ratio is seen. It is
extremely difficult to rationalize this behavior in terms
of the standard GE model. This model would predict (i)
a strong S-driving force from GE(O4′-C4′-C3′-F3′)
counteracted by the much weaker N-driving O4′-C1′-
C2′-Cl2′ and (ii) mutual cancellation of the cis GEs of
B-C1′-C2′-Cl2′. The S-driving GE is augmented by the
additional S drive from the AE. In all, the standard view
predicts a highly predominant S population that is
(almost) independent of the nature of the nucleobase. The
AP model offers a satisfying rationalization: (i) The
strongly S-driving AP[F3′-H2′] is counteracted by two
N-driving APs induced by chlorine, AP[Cl2′-H3′] and
AP[Cl2′-H1′]. Since the relative electronegativity ∆ø )
øX - øH of Cl (0.99) is about half that of F (1.9) on the
Mullay scale,59b the opposing drives of the two halogens
will mutually cancel. (ii) There remains a weaker S-
driving AP[F3′-C5′]. (iii) The S-driving AP[B-H2′] is
probably weak, and the situation is dominated by the
S-driving AE of the base, thus qualitatively explaining
the base dependency of the population ratio. This behav-
ior runs counter to expectations based on the strength
of the AE displayed by 2′-deoxy-R-D-nucleosides (R-D-dNs)
wherein the S form is relatively more stabilized by the
nucleobase in the sequence thymine-1-yl < adenin-9-yl
< cytosin-1-yl.50 Perhaps unfortunately, the conforma-
tional behavior and thermodynamics of a series of R-D-
rNs, to the best of our knowledge, seem not to have been
investigated, and extrapolation from R-D-dNs to R-D-rNs
need not be necessarily correct.
Compared to the F3′-substituted â-D-nucleosides dis-
cussed above, a more outspoken behavioral differentia-
tion is displayed by the methyl 2,3-dideoxy-3-fluoro-2-
(X-substituted)-â-D-ribofuranosides (X ) H, Cl, and F;
Table 2), where we find a progressive shift toward the N
form when the electronegativity of X increases: 31-36%
N for X ) H (30, 31), 59-60% N for X ) Cl (2, 4, 5), and
82-85% N for X ) F (28, 29). This trend appears to run
counter to predictions based on the conventional GE
approach. For example, when X ) F, GE(F2′-C2′-C1′-
O4′) in the N form should balance GE(F3′-C3′-C4′-O4′)
in the S form, but there remains the strongly S-driving
GE(F2′-C2′-C1′-OMe), which counteracts the N-driv-
ing AE(OMe). The data can be rationalized by assuming
that AE(OMe) . GE(F2′-C2′-C1′-OMe), but this as-
sumption does not fit the predominance of the S form
when X ) H, where obviously GE(F3′-C3′-C4′-O4′) .
AE(OMe). The experimental trend is well compatible with
our proposed AP description in conjunction with the
anomeric effect. First, one recognizes that two important
APs, AP[X2′-OMe] in the N form and AP[X2′-O4′] in
the S form, balance for each given X substituent. Second,
in the special case that X ) F the effects of all strongly
driving APs cancel, and only three weaker ones remain:
AP[F3′-O4′] in the N form vs AP[OMe-C3′] plus AP-
[F3′-C5′] in the S form. A partial cancellation occurs,
the AE(OMe) assumes the predominant role, and the N
form is favored (∆G ) +4.0 to +4.5 kJ ‚mol-1). In the case
that X ) H the net result of various cancellations is that
the S-driving AP[F3′-H2′], aided by AP[OMe-C3′],
dominates over the N-driving AE(OMe) (∆G ) -1.5 to
-2 kJ ‚mol-1). With a 2′-substituent of moderate elec-
tronegativity (X ) Cl) it is no surprise to find a ∆G value
between the two extremes (+0.9 to +1.1 kJ ‚mol-1).
The strong AE(OMe), due to the high electronegativity
of OMe combined with its relatively small steric size
compared to that of a nucleobase, also helps to explain
the peculiar pseudorotational behavior of the first mem-
ber of this series (X ) H): PN ≈ -60° and PS ≈ 225°. By
contrast, normal 2′-deoxy-â-D-ribonucleosides prefer to
occupy the ranges PN ) 0°-18° and PS ) 144°-180°. It
is generally agreed that for maximum n f σ* interaction
the orbital of the lone pair donor should be aligned with
the antibonding orbital of the acceptor bond. The question
now pertains to the best description of the hybridization
of the lone pairs of an ether oxygen atom. In the course
of time two different models were proposed: the sp3
model and the sp2 (π-like) model. Although in descriptions
of the anomeric effect the sp3 model was originally
favored,56 it now appears preferable65 to use the π-like
model wherein the p lone pair is positioned perpendicular
to the C1′-O4′-C4′ plane. Consideration of the Newman
projections along the C1′-O4′ bond shows that in â-D-
ribosides a pseudorotation of PN toward negative values
(PN ) -36° to -54°) brings about a perfect antiparallel
alignment between the p-type lone pair of O4′ and the
C1′-O1′ bond (181°-187°, respectively). Even more
interesting is the observation that pseudorotation of PS
beyond 180° (high S, PS ) 216°-234°) improves the same
alignment (162°-173°), suggesting that the high-S form
of â-D-ribofuranosides is also stabilized by a certain
amount of anomeric interaction. By contrast, the geo-
metrical situation in R-D-ribosides is completely different.
The observed pseudorotation of the N form toward
negative PN appears to be related to an increasingly
improving alignment of AP[OMe-C4′] (157° for PN
)
-54°); in the S form one predicts the maximum favorable
anomeric alignment of 180° to occur near PS ) 144°, and
this corresponds with the results of the present PSEUROT
analyses.
Finally, the conformational populations of the methyl
R-D-ribofuranosides merit a brief discussion. Remarkably,
for X2′ ) Cl (6) the PSEUROT calculations yield an
unusually wide range of satisfactory (low rms) solutions
of the least-squares problem. Representative examples
are shown in Table 2 (6I and 6II). It is interesting to see
that different values of ∆Geff hardly affect the resulting
geometrical parameters but yield different populations.
For both X2′ ) Cl (6) and X2′ ) OH (32) the AP[F3′-
H2′] cooperates with the AE(OMe) to drive the sugar ring
to adopt the S form. For X2′ ) OH the N-driving AP-
[O2′-H1′′] balances the S-driving AP[OMe-H2′]. No
such balancing is expected for X2′ ) Cl, and one predicts
a strong predominance of the S-type conformer.
Con clu sion s
We have shown that the R-arranged chlorine substitu-
ent at the C2 carbon of methyl pentofuranosides can be
used as the anomeric control factor in the glycosylation
(65) Cosse-Barbie, A.; Watson, D. G.; Dubois, J . E. Tetrahedron Lett.
1989, 30, 163-166.
J . Org. Chem, Vol. 68, No. 15, 2003 5905

Mikhailopulo et al.
mmol) in DMSO (15 mL) was added LiCl (0.43 g, 10 mmol),
and the mixture was stirred for 2 h at 200 °C (bath temper-
ature). After being cooled to ambient temperature, the mixture
was partitioned between CHCl3 (50 mL) and H2O (30 mL). The
organic layer was separated, washed with H2O (4 × 30 mL),
dried, and evaporated. The residue was chromatographed on
a silica gel column (3 × 30 cm) using a linear gradient of Et2O
in hexane (10% f 50%, v/v; 1.0 L) to yield compound 2 (0.15
g; 55%) as a syrup, Rf 0.72 (A), and the starting material 1
(50 mg; 12%).
Meth yl 5-O-Ben zoyl-2-ch lor o-2,3-d id eoxy-3-flu or o-â-D-
r ibofu r a n osid e (5). Meth od A. F r om 2. To a solution of 2
(0.5 g, 1.8 mmol) in EtOH (50 mL) was added 5% Pd/C (0.5 g),
and the mixture was stirred in a H2 atmosphere for 18 h. The
catalyst was filtered off and washed with EtOH (3 × 20 mL).
The combined filtrates were evaporated, and the residue was
coevaporated with toluene (2 × 20 mL) to afford the oily
residue of 4 [0.3 g, 1.6 mmol; 89%; Rf 0.26 (A)]. The residue
was dissolved in anhydrous pyridine (10 mL), BzCl (0.2 mL,
0.25 g, 1.76 mmol) was added, and the mixture was stirred
for 18 h. The mixture was diluted with CHCl3(50 mL), washed
with H2O (30 mL), aqueous NaHCO3 (30 mL), H2O (30 mL), 1
N H2SO4 (30 mL), and H2O (30 mL), dried, evaporated, and
chromatographed as described above to give compound 5 (0.42
g; 92%) as a syrup: Rf 0.65 (A).
reaction. Subsequent efficient removal of the chlorine
atom under the action of Bu3SnH/AIBN afforded 2′-
deoxynucleosides. Employing methyl 5-O-benzoyl-2-chloro-
2,3-dideoxy-3-fluoro-â-D-ribofuranoside (5) as the univer-
sal glycosylating agent, the 2′,3′-dideoxy-3′-fluoro-â-D-
erythro-pentofuranosyl nucleosides of thymine (FLT),
cytosine (FLC), and adenine (FLA) were synthesized
along with the corresponding R-anomers RFLT, RFLC,
and RFLA as byproducts. Dehalogenation of 2′-chloro-
2′,3′-dideoxy-3′-fluoro-â-D-ribofuranosides of thymine (10b)
and adenine (11b) with 1 M MeONa/MeOH resulted in
exclusive HF elimination, giving rise to the respective
vinyl chlorides 16 and 19 as the principal products.
Similar easy HF elimination was also observed in the
cases of the blocked anomeric adenine R- and â-nucleo-
sides 8a ,b and of 2′-azido-2′,3′-dideoxy-3′-fluoro-â-D-ri-
bofuranosides of thymine (17) and adenine (20). It is
suggested that this behavior is governed by the antipar-
allel configuration of the C3′-F and the C2′-H bonds in
the predominant S conformation of the pentofuranose
ring in the above compounds.
The conformational behavior of the 3′-fluorinated fura-
3
3
noside rings was explored on the basis of J HH and J HF
with the aid of the PSEUROT6.3 program.47 The predic-
tive power of the standard model of the GE54a was
compared to that of an alternative model proposed by
Brunck and Weinhold,57a which invokes maximum AP σ
T σ* stabilization when the donating bond is the least
polar one and the acceptor orbital is at the most polarized
bond. The 2′,3′-dideoxy-3′-fluoro-2′-(X-substituted)-â-D-
ribonucleosides (X2′ ) H, Cl, N3, NH2, and OH) display
a strong preference for the S conformation (g94% S),
independent of the nature of the nucleobase. When X2′
) F (27), this preference drops to 80% S. Both models
tested appear to explain these data well, at least quali-
tatively. Matters are different in the corresponding R-D-
nucleoside series 10a , 11a , and 12a (X2′ ) Cl), which
displays a remarkable base-dependent progression in the
population of the N form in going from Ade to Thy to
Cyt. Here the AP model offers a satisfying rationalization,
whereas the standard view would predict a strong
predominance of the S form in all compounds. Similarly,
the conformational behavior of the corresponding sub-
stituted methyl R- and â-D-ribofuranosides is more com-
patible with predictions from the AP model than it is with
those from the standard GE model. However, further
research is needed before a final verdict concerning the
relative merits of the two competing views can be passed.
Meth od B. F r om 3. Tosylate 3 was prepared from 1 in 71%
yield.35 To a solution of 3 (2.05 g, 4.83 mmol) in DMSO (60
mL) was added LiCl (2.04 g, 48.2 mmol), and the mixture was
stirred for 2 h at 200 °C (bath temperature). The workup and
purification, performed as described above for 2, yielded 5 (1.0
g; 72%) and the starting tosylate 3 (0.33 g; 16%).
1-(2-Ch lor o-2,3-d id eoxy-3-flu or o-â-D-r ib ofu r a n osyl)-
th ym in e (10b) a n d Its r-An om er (10a ). A solution of 5 (0.42
g, 1.45 mmol), the bis(trimethylsilyl) derivative of thymine
[obtained from 0.37 g (2.9 mmol) of thymine and (TMS)Tfl (0.79
mL, 0.97 g, 4.35 mmol)] in anhydrous CH3CN (30 mL) was
refluxed for 5 h and then cooled to room temperature. The
reaction mixture was diluted with CHCl3 (100 mL) and washed
with aqueous NaHCO3 (100 mL). The water layer was washed
with CHCl3 (2 × 50 mL). The organic extracts were combined,
washed with H2O (50 mL), dried, and evaporated. The residue
was chromatographed on a silica gel column (3 × 30 cm) using
a linear gradient of EtOAc in hexane (10% f 90%, v/v; 2.0 L)
to yield a mixture of the starting material 5 and its R-anomer
6 (80 mg; 19%) and a mixture of the nucleosides 7a ,b (0.27 g;
49%). Rechromatography of a mixture of glycosides 5 and 6
gave the individual â-anomer 5 (30 mg) and the R-anomer 6
[35 mg; Rf 0.62 (A)]. The mixture of 7a ,b (0.24 g, 0.62 mmol)
in methanol, saturated with ammonia at 0 °C (25 mL), was
stirred for 24 h and evaporated. The residue was chromato-
graphed on a silica gel column (3 × 30 cm) using a linear
gradient of MeOH in CHCl3 (0% f 10%, v/v; 1.5 L) to yield in
order of elution 10b (95 mg; 55%) and its R-anomer 10a (48
mg; 28%). Data for 10b: Rf 0.36 (B); mp 222-223 °C (from
EtOAc-hexane); UV λmax 263 nm (ꢀ 9700). Anal. Calcd for
Exp er im en ta l Section
C
10H12ClFN2O4 (278.66): C, 43.10; H, 4.34. Found: C, 43,22;
H, 4.62. Data for 10a : Rf 0.30 (B); mp 206-209 °C (from
EtOAc-hexane); UV λmax 265 nm (ꢀ 9800). Anal. Found: C,
43.18; H, 4.54.
Gen er a l P r oced u r es. Melting points are uncorrected. The
UV spectra were recorded in ethanol, except where otherwise
indicated. The 1H and 13C NMR spectra were measured at
200.13 and 50.325 MHz, respectively, at 23 °C with tetra-
methylsilane as an internal standard. Standard Silufol UV254
and silica gel 60F254 plates were used for thin-layer chroma-
tography (TLC) of sugars and nucleosides, respectively; solvent
systems used were hexane-diethyl ether (1:1) (A), CHCl3-Me-
OH (9:1) (B), and CHCl3-MeOH (4:1) (C). Column chromatog-
raphy was performed on a silica gel L 40/100 µm column.
Freshly distilled anhydrous DMSO was used throughout this
work. Except otherwise indicated, the reactions were carried
out at room temperature. The solutions of compounds in
organic solvents were dried with anhydrous Na2SO4 for 4 h.
Met h yl 5-O-Ben zyl-2-ch lor o-2,3-d id eoxy-3-flu or o-â-D-
r ibofu r a n osid e (2). To a stirred solution of 135 (0.41 g, 1.0
2′-Ch lor o-2′,3′-d id eoxy-3′-flu or oa d en osin e (11b) a n d
Its r-An om er (11a ). A solution of 5 (0.4 g, 1.38 mmol), the
bis(trimethylsilyl) derivative of N6-benzoyladenine [obtained
from 0.66 g (2.76 mmol) of N6-benzoyladenine], and SnCl4 (0.8
mL, 1.8 g, 6.9 mmol) in an anhydrous acetonitrile/1,2-dichlo-
roethane mixture (45 mL; 2:1, v/v) was refluxed for 5 h. After
standard workup and chromatography on a silica gel column
(3 × 30 cm) using a linear gradient of EtOAc in hexane (10%
f 90%, v/v; 2.0 L), a mixture of 5 and 6 (30 mg; 7.5%) and
individual 8b (0.45 g; 66%) and its R-anomer 8a (60 mg; 9%)
were isolated. Standard deprotection of 8b (0.25 g, 0.5 mmol)
followed by column chromatography [silica gel column, 2 ×
32 cm; a linear gradient of MeOH in CHCl3 (0% f 10%, v/v;
5906 J . Org. Chem., Vol. 68, No. 15, 2003

2′-Chloro-2′,3′-dideoxy-3′-fluoro-D-ribonucleosides
1
1.5 L)] and crystallization from EtOH yielded 11b (0.12 g;
83%): Rf 0.3 (B); mp 236 °C; UV λmax 260 nm (ꢀ 14700). Anal.
Calcd for C10H11ClFN5O2 (287.68): C, 41.75; H, 3.85. Found:
C, 41.48, H, 4.07. In a similar way, starting from 8a (60 mg,
0.12 mmol), deblocked nucleoside 11a was obtained (27 mg;
78%): Rf 0.2 (B); mp 105-106 °C (from EtOH); UV λmax 260
nm (ꢀ 14300). Anal. Found: C, 41.37; H, 4.02.
respects (mp and UV and H and 13C NMR spectroscopy) with
the corresponding authentic samples.24
Meth od B. To a solution of 11b (15 mg, 0.052 mmol) in a
dioxane/MeOH mixture (3 mL, 1:1, v/v) were added Bu3SnH
(0.07 mL, 76 mg, 0.26 mmol) and AIBN (3 mg, 0.018 mmol),
and the reaction mixture was refluxed for 15 h. After standard
workup and chromatography, the â-fluoride 14b (11 mg; 84%)
was obtained.
2′-Ch lor o-2′,3′-d id eoxy-3′-flu or ocytid in e (12b) a n d Its
r-An om er (12a ). A solution of 5 (0.51 g, 1.76 mmol), the bis-
(trimethylsilyl) derivative of N4-benzoylcytosine [obtained from
0.76 g (3.52 mmol) of N4-benzoylcytosine], and SnCl4 (0.67 mL,
1.37 g, 5.28 mmol) in anhydrous CH3CN (50 mL) was refluxed
for 5 h. After standard workup and chromatography on a silica
gel column (3 × 30 cm) using a linear gradient of EtOAc in
hexane (10% f 90%, v/v; 2.0 L), a mixture of 5 and 6 (90 mg;
18%) and individual 9b (0.47 g; 56%) and its R-anomer 9a (70
mg; 8%) were isolated in order of elution. Standard deprotec-
tion of 9b (0.24 g, 0.51 mmol) followed by column chromatog-
raphy [silica gel column, 2 × 32 cm; a linear gradient of MeOH
in CHCl3 (0% f 20%, v/v; 1.5 L)] and crystallization from
EtOH yielded 12b (0.105 g; 78%): Rf 0.48 (C); mp 111-112
°C; UV (pH 7.0) λmax 243 nm (ꢀ 8800), 265 nm (shoulder); (pH
2.0) λmax 280 nm (ꢀ 13500). Anal. Calcd for C9H11ClFN3O3
(263.65): C, 41.00; H, 4.21. Found: C, 41.23; H, 4.58. In a
similar way, starting from 9a (70 mg, 0.15 mmol), deblocked
nucleoside 12a was obtained (31 mg; 80%): Rf 0.45 (C); mp
223-225 °C (from EtOH); UV (pH 7.0) λmax 241 nm (ꢀ 8400),
273 nm (8700); (pH 2.0) λmax 282 nm (ꢀ 14300). Anal. Found:
C, 41.27; H, 4.60.
2′,3′-Did eoxy-3′-flu or ocytid in e (15b) a n d Its r-An om er
(15a ). Meth od A. To a solution of 9b (0.1 g, 0.21 mmol) in
anhydrous dioxane (10 mL) were added Bu3SnH (0.34 mL, 0.39
g, 1.26 mmol) and AIBN (10 mg, 0.06 mmol), and the reaction
mixture was refluxed for 16 h and evaporated. The residue
was treated with methanol, saturated with ammonia at 0 °C
(10 mL) for 48 h, and evaporated. The residue was triturated
with pentane, and the precipitate was filtered off, washed with
pentane, and purified by silica gel column chromatography as
described above for 12b to afford the â-fluoride 15b [15 mg;
31%; Rf 0.42 (C)].
Meth od B. To a solution of 12b (16 mg, 0.061 mmol) in an
anhydrous dioxane/MeOH mixture (3.0 mL; 1:1, v/v) were
added Bu3SnH (0.06 mL, 88 mg, 0.3 mmol) and AIBN (3 mg,
0.018 mmol), and the reaction mixture was refluxed for 15 h.
After standard workup and chromatography, the â-fluoride
15b (12 mg; 86%) was obtained. Similarly, starting from 12a
(15 mg, 0.057 mmol), the R-fluoride 15a [10 mg; 77%; Rf 0.41
(C)] was obtained. Compounds 15b and 15a were identical in
all respects (mp and UV and 1H and 13C NMR spectroscopy)
with the corresponding authentic samples.12
3′-Deoxy-3′-flu or oth ym id in e (13b) a n d Its r-An om er
(13a ). Meth od A. To a solution of a mixture of nucleosides
7a ,b (76 mg, 0.2 mmol) in anhydrous dioxane (10 mL) were
added Bu3SnH (0.31 mL, 0.35 g; 1.2 mmol) and AIBN (10 mg,
0.06 mmol), and the reaction mixture was refluxed for 2 h and
evaporated. The residue was treated with methanol, saturated
with ammonia at 0 °C (25 mL) for 24 h, and evaporated. The
residue was portioned between MeOH (30 mL) and pentane
(10 mL). The methanol layer was separated and evaporated,
and the residue was chromatographed on a silica gel column
(2 × 32 cm) using a linear gradient of MeOH in CHCl3 (0% f
10%, v/v; 1.5 L) to afford nucleoside 13b [27 mg; 55%; Rf 0.3
(B)] and its R-anomer 13a [13 mg; 27%; Rf 0.25 (B)]. Both
compounds 13b and 13a were identical in all respects (mp and
1-(2-Ch lor o-2,3-d id eoxy-â-D-glycer o-p en t-2-en ofu r a n o-
syl)th ym in e (16). To a solution of 10b (60 mg, 0.22 mmol) in
CH3OH (5 mL) was added 0.45 mL of a 1 M solution of CH3-
ONa in CH3OH, and the reaction mixture was refluxed for 5
h. After cooling, the reaction mixture was neutralized with 1
N CH3COOH in CH3OH and evaporated, and the residue was
chromatographed on a silica gel column (2 × 32 cm) using a
linear gradient of MeOH in CHCl3 (0% f 10%, v/v; 1.5 L) to
give in order of elution unconsumed starting nucleoside 10b
(10 mg; 17%) and olefin 16 (26 mg; 46%): Rf 0.28 (B); mp 192-
195 °C (from EtOAc-hexane); UV λmax 263 nm (ꢀ 9500). Anal.
Calcd for C10H11ClN2O4 (258.66): C, 46.44; H, 4.28; Cl, 13.71.
Found: C, 46.71; H, 4.10; Cl, 14.07.
1-(2-Azido-2,3-dideoxy-â-D-glycer o-pen t-2-en ofu r an osyl)-
th ym in e (18). To a solution of 1735 (175 mg, 0.61 mmol) in
CH3OH (15 mL) was added 1.2 mL of 1 M CH3ONa in CH3-
OH, and the reaction mixture was refluxed for 5 h. After
standard workup and chromatography, unconsumed starting
nucleoside 17 (58 mg; 33%) and 18 (76 mg; 47%) were isolated.
Data for 18: Rf 0.29 (B); mp 143-145 °C (from EtOAc-
1
UV and H and 13C NMR spectroscopy) with the corresponding
authentic samples.12,18
Meth od B. To a solution of deblocked nucleoside 10b (56
mg, 0.2 mmol) in anhydrous dioxane (10 mL) were added Bu3-
SnH (0.26 mL, 0.29 g; 1.0 mmol) and AIBN (10 mg, 0.06 mmol),
and the reaction mixture was refluxed for 2 h and evaporated.
The residue was triturated with pentane, and the precipitate
was filtered off, washed with pentane, and purified by silica
gel column chromatography as described above to afford the
â-fluoride 13b (44 mg; 90%).
Meth od C. To a solution of 10b (28 mg, 0.1 mmol) in a
mixture of dioxane/MeOH (5 mL, 1:1; v/v) were added Bu3-
SnH (0.13 mL, 0.145 g; 0.5 mmol) and AIBN (5 mg, 0.03 mmol),
and the reaction mixture was refluxed for 15 h. After workup
and chromatography as described in method A or B, the
â-fluoride 13b (22 mg; 90%) was obtained.
2′,3′-Did eoxy-3′-flu or oa d en osin e (14b) a n d Its r-An o-
m er (14a ). A. To a solution of 8b (0.3 g, 0.6 mmol) in
anhydrous dioxane (30 mL) were added Bu3SnH (0.95 mL, 1.05
g, 3.6 mmol) and AIBN (30 mg, 0.18 mmol), and the reaction
mixture was refluxed for 6 h and evaporated. The residue was
treated with methanol, saturated with ammonia at 0 °C (25
mL) for 48 h, and evaporated. The residue was triturated with
pentane, and the precipitate was filtered off, washed with
pentane, and purified by silica gel column chromatography as
described above to afford the â-fluoride 14b [120 mg; 79%; Rf
0.25 (B)]. In a similar way, starting from 8a (60 mg, 0.12
mmol), the R-fluoride 14a [25 mg; 82%; Rf 0.15 (B)] was
obtained. Compounds 14b and 14a were identical in all
hexane); UV λmax 263 nm (ꢀ 9400); IR (KBr) ν(N3) 2120 cm-1
.
Anal. Calcd for
C10H11N5O4 (265.22): C, 45.29; H, 4.18.
1
Found: C, 45.61; H, 3.96. The H and 13C NMR spectra are in
accord with those measured for the same compound previously
synthesized by an alternative procedure40 and obtained as an
amorphous powder.
9-(2-Ch lor o-2,3-d id eoxy-â-D-glycer o-p en t-2-en ofu r a n o-
syl)a d en in e (19) a n d Its r-An om er (22). Meth od A. To a
solution of 8b (80 mg, 0.16 mmol) in CH3OH (10 mL) was
added 0.24 mL of 1 M CH3ONa in CH3OH, and the reaction
mixture was refluxed for 4 h. Standard workup and chroma-
tography gave debenzoylated nucleoside 11b (5 mg; 11%) and
19 (35 mg; 82%): Rf 0.24 (B); mp 240 °C dec (from EtOH); UV
λ
max 260 nm (ꢀ 14800). Anal. Calcd for C10H10ClN5O2 (267.67):
C, 44.87; H, 3.76; Cl, 13.25. Found: C, 45.20; H, 3.48; Cl, 13.58.
Meth od B. Similarly, starting from 11b (50 mg, 0.17 mmol),
unconsumed nucleoside 11b (7 mg; 12%) and 19 (36 mg; 77%)
were isolated.
Meth od C. In a similar way, treatment of 8a (75 mg, 0.15
mmol) under reflux for 1 h gave, after standard workup and
chromatography, compound 22 (35 mg; 86%): Rf 0.13 (B); mp
173-175 °C (from EtOH); UV λmax 260 nm (ꢀ 14300). Anal.
Found: C, 44.91; H, 3.97; Cl, 13.05.
J . Org. Chem, Vol. 68, No. 15, 2003 5907

Mikhailopulo et al.
9-(2-Azid o-2,3-d id eoxy-â-D-glycer o-p en t-2-en ofu r a n osy-
l)a d en in e (21) was obtained, as described above for compound
19, starting from 50 mg (0.17 mmol) of 2035 to give 32 mg (69%)
as a slightly yellow amorphous powder: Rf 0.25 (B); UV λmax
259 nm (ꢀ 14500); IR (KBr) ν(N3) 2120 cm-1. Anal. Calcd for
Su p p or tin g In for m a tion Ava ila ble: Synthesis and char-
acterization of 27-31 and 1H and 13C NMR data (chemical
shifts and coupling constants) of a series of selected com-
pounds. This material is available free of charge via the
Internet at http://pubs.acs.org.
C
10H10N8O2 (274.24): C, 43.80; H, 3.67. Found: C, 44.07; H,
3.41.
J O0340859
5908 J . Org. Chem., Vol. 68, No. 15, 2003